drug calculation assignment

Drug Dosage Calculations NCLEX Practice Questions (100+ Items)

Drug Dosage Calculations Nursing Test Banks for NCLEX RN

Welcome to your NCLEX reviewer for nursing drug calculations! In this nursing test bank , practice dosage calculation problems to measure your competence in nursing math. As a nurse , you must accurately and precisely calculate medication dosages to provide safe and effective nursing care. This quiz aims to help students and registered nurses alike grasp and master the concepts of medication calculation.

Drug Dosage Calculation Practice Quiz

In this section are the practice problems and questions for nursing dosage calculations. This nursing test bank set includes 100+ questions. Included topics are dosage calculation, metric conversions, unit conversions, parenteral medications, and fluid input and output. As you can tell, this NCLEX practice exam requires tons of calculations, so get your calculators ready!

Remember to answer these questions at your own pace, and don’t forget to read the rationales! Don’t be discouraged if you have incorrect answers. You are here to learn! Make sense of the rationales and review the drug dosage calculations study guide below.

Quiz Guidelines

Before you start, here are some examination guidelines and reminders you must read:

Practice Exams : Engage with our Practice Exams to hone your skills in a supportive, low-pressure environment. These exams provide immediate feedback and explanations, helping you grasp core concepts, identify improvement areas, and build confidence in your knowledge and abilities.
You’re given 2 minutes per item.
For Challenge Exams, click on the “Start Quiz” button to start the quiz.
Complete the quiz : Ensure that you answer the entire quiz. Only after you’ve answered every item will the score and rationales be shown.
Learn from the rationales : After each quiz, click on the “View Questions” button to understand the explanation for each answer.
Free access : Guess what? Our test banks are 100% FREE. Skip the hassle – no sign-ups or registrations here. A sincere promise from Nurseslabs: we have not and won’t ever request your credit card details or personal info for our practice questions. We’re dedicated to keeping this service accessible and cost-free, especially for our amazing students and nurses. So, take the leap and elevate your career hassle-free!
Share your thoughts : We’d love your feedback, scores, and questions! Please share them in the comments below.

Quizzes included in this guide are:

Drug Calculations Reviewer for Nurses

This is your study guide to help you refresh or review what you know about drug dosage calculations, including tips on answering them.

NCLEX Tips for Dosage Calculation Questions

The fill-in-the-blank question format is usually used for medication calculation, IV flow rate calculation, or determining the intake-output of a client. In this question format, you’ll be asked to perform a calculation and type in your answer in the blank space provided.
Always follow the specific directions as noted on the screen.
The unit of measure you need for your final answer is always given.
There will be an on-screen calculator on the computer for you to use.
Do not put any words, units of measurements, commas, or spaces with your answer, type only the number. Only the number goes into the box. Rounding an answer should be done at the end of the calculation or as what the question specified, and if necessary, type in the decimal point.

Nursing Responsibilities for Medication Administration

Right Drug. The first right of drug administration is to check and verify if it’s the right name and form. Beware of look-alike and sound-alike medication names. Misreading medication names that look similar is a common mistake. These look-alike medication names may also sound alike and can lead to errors associated with verbal prescriptions. Check out The Joint Commission’s list of look-alike/sound-alike drugs .
Right Patient . Ask the name of the client and check his/her ID band before giving the medication. Even if you know that patient’s name, you still need to ask just to verify.
Right Dose . Check the medication sheet and the doctor’s order before medicating. Be aware of the difference between an adult and a pediatric dose.
Right Route . Check and verify the order (i.e., per orem, IV, SQ, IM)
Right Time and Frequency. Check the order for when it would be given and when was the last time it was given.
Right Documentation . Make sure to write the time and any remarks on the chart correctly.
Right History and Assessment. Secure a copy of the client’s history to drug interactions and allergies.
Right Drug Approach and Right to Refuse . Give the client enough autonomy to refuse the medication after thoroughly explaining the effects.
Right Drug-Drug Interaction and Evaluation. Review any medications previously given or the diet of the patient that can yield a bad interaction to the drug to be given. Check also the expiry date of the medication being given.
Right Education and Information. Provide enough knowledge to the patient of what drug he/she would be taking and what are the expected therapeutic and side effects.

Systems of Measurement

There are three systems of measurement used in nursing: the metric system, the apothecaries’ system, and household system.
The most widely used international system of measurement.
The basic units of metric measures are the gram (weight) , meter (length or distance) , and liter (volume) .
It is a decimal-based system that is logically organized into units of 10. Basic units are multiplied or divided by 10 to form secondary units.
The apothecaries’ system is one of the oldest systems of measurement, older than the metric system and is considered to be out of date.
The basic units used in this system are the grain (gr) for weight, minim for volume, ounce, and pound. All of which are seldomly used in the clinical setting.
Quantities in the apothecaries’ system are often expressed by lowercase Roman numerals when the unit of measure is abbreviated. And the unit of measure precedes the quantity. Quantities less than 1 are expressed as fractions. Examples: “gr ii”, “gr ¼ ”
And yes, it can be confusing therefore use the metric system instead to avoid medication errors .
Household system measures may be used when more accurate systems of measure are not required.
Included units are drops, teaspoons, tablespoons, cups, pint, and glasses.
The milliequivalent is an expression of the number of grams of a medication contained in 1 milligram of a solution.
Examples: the measure of serum sodium , serum potassium , and sodium bicarbonate is given in milliequivalents.
Unit measures a medication in terms of its action, not its physical weight.
When documenting, do not write “U” for unit, rather spell it as “unit” as it is often mistaken as “0”.
Examples: Insulin , penicillin , and heparin sodium are measured in units.

Converting Units of Weight and Measure

For drug dosages, the metric units used are the gram (g), milligram (mg), and microgram (mcg) . For volume units milliliters (mL) and liters (L).
It is simple to compute for equivalents using the metric system. It can be done by dividing or multiplying; or by moving the decimal point three places to the left or right.
Do not use a “trailing zero” after the decimal point when the dosage is expressed as a whole number. For example, if the dosage is 2m mg, do not insert a decimal point or the trailing zero as this could be mistaken for “20” if the decimal point is not seen.
On the other hand, do not leave a “naked” decimal point. If a number begins with a decimal, it should be written with a zero and a decimal point before it. For example, if the dosage is 2/10 of a milligram, it should be written as 0.2 mg. It could be mistaken for 2 instead of 0.2.
Household and metric measures are equivalent and not equal measures.
Conversions to equivalent measures between systems is necessary when a medication prescription is written in one system but the medication label is stated in another.
Medications are not always prescribed and prepared in the same system of measurement; therefore conversion of units from one system to another is necessary.
Common conversions in the healthcare setting include pound to kilograms, milligrams to grains, minims to drops.

Methods for Drug Dosage Calculations

The commonly used formula for calculating drug dosages.
D = Desired dose or dose ordered by the primary care provider.
H = dose on hand or dose on the label of bottle, vial, ampule.
V = vehicle or the form in which the drug comes (i.e., tablet or liquid).

STANDARD FORMULA Formula = \frac{Desired (D) \times Vehicle (V) }{On\ Hand (H)} = amount \ to \ administer

Considered as the oldest method used for drug calcluation problems.
For the equation, the known quantities are on the left side, while the desired dose and the unknown amount to administer are on the right side.
X = amount to administer
Once the equation is set up, multiply the extremes (H and x ) and the means (V and D). Then solve for x .

RATIO AND PROPORTION METHOD H : V = D : x

A method similar to ratio and proportion but expressed as fractions.

FRACTIONAL EQUATION METHOD \frac{H}{V}= \frac{D}{x}

Intake and output (I&O) measurement and recording is usually done to monitor a client’s fluid and electrolyte balance during a 24-hour period.
Intake and output is done for patients with increased risk for fluid and electrolyte imbalance (e.g., heart failure , kidney failure).
Unit used in measurement of I&O is milliliter (mL) .
Oral fluids (e.g., water, juice, milk, soup, water taken with medication).
Liquid foods at room temperature (e.g., ice cream, gelatin, custard).
Tube feedings including the water used for flushes.
Parenteral fluids
Blood products
IV medications
Urinary output
Liquid feces
Tube drainage
Wound and fistula drainage
Measurement of fluid input and output are totaled at the end of the shift and documented in the patient’s chart.
Determine if fluid intake and fluid output are proportional. When there is a significant discrepancy between intake and output, report to the primary care provider.

Recommended Resources

Recommended books and resources for your NCLEX success:

Disclosure: Included below are affiliate links from Amazon at no additional cost from you. We may earn a small commission from your purchase. For more information, check out our privacy policy .

Saunders Comprehensive Review for the NCLEX-RN Saunders Comprehensive Review for the NCLEX-RN Examination is often referred to as the best nursing exam review book ever. More than 5,700 practice questions are available in the text. Detailed test-taking strategies are provided for each question, with hints for analyzing and uncovering the correct answer option.

Strategies for Student Success on the Next Generation NCLEX® (NGN) Test Items Next Generation NCLEX®-style practice questions of all types are illustrated through stand-alone case studies and unfolding case studies. NCSBN Clinical Judgment Measurement Model (NCJMM) is included throughout with case scenarios that integrate the six clinical judgment cognitive skills.

Saunders Q & A Review for the NCLEX-RN® Examination This edition contains over 6,000 practice questions with each question containing a test-taking strategy and justifications for correct and incorrect answers to enhance review. Questions are organized according to the most recent NCLEX-RN test blueprint Client Needs and Integrated Processes. Questions are written at higher cognitive levels (applying, analyzing, synthesizing, evaluating, and creating) than those on the test itself.

NCLEX-RN Prep Plus by Kaplan The NCLEX-RN Prep Plus from Kaplan employs expert critical thinking techniques and targeted sample questions. This edition identifies seven types of NGN questions and explains in detail how to approach and answer each type. In addition, it provides 10 critical thinking pathways for analyzing exam questions.

Illustrated Study Guide for the NCLEX-RN® Exam The 10th edition of the Illustrated Study Guide for the NCLEX-RN Exam, 10th Edition. This study guide gives you a robust, visual, less-intimidating way to remember key facts. 2,500 review questions are now included on the Evolve companion website. 25 additional illustrations and mnemonics make the book more appealing than ever.

NCLEX RN Examination Prep Flashcards (2023 Edition) NCLEX RN Exam Review FlashCards Study Guide with Practice Test Questions [Full-Color Cards] from Test Prep Books. These flashcards are ready for use, allowing you to begin studying immediately. Each flash card is color-coded for easy subject identification.

31 thoughts on “Drug Dosage Calculations NCLEX Practice Questions (100+ Items)”

Part 1: 13/15 Part 2: 33/40 Part 3: 43/50 Part 4: 9/10

Challenging but fun!

Let’s elevate the discourse. Petty, negative remarks are unnecessary.

I agree, negative remarks are unnecessary, especially when the time has been taken to make this information available to us.

Hey, you can always correct/point out people’s mistakes politely, no need to be an ass about it. Being a nurse and having a bachelors degree does not mean one has to be perfect (unless you’re perfect? lol). I can imagine what kind of ‘nurse’ you are/will be. Your lack of manners makes me cringe.

I had my first experience working with RNs through the covid times and the person I worked with and trained me was like that he wanted or expected me to think like him and do everything like him and if I would ask him a question to confirm he would say things like “didn’t I explain that already or something like a smart allic ” trust me I am very proud not to have punched him all of these times but he was harmless in nursing there are just those people that don’t think about others and just expect you’re like them or if your not your below them which is unfortunate!

I learned how not to be and how to act I would even help the new RNs once I was concerned not new and I would be determined not to treat anyone how I was treated I don’t think it was A RN thing it was either you on his level or not so after I was comfortable I started going off on him bickering back and forth but he had to know I am not the one and I was new so I let it slide but don’t make those mistake anymore! he would sabotage me I have to admit he did it a way that no one knew very smart which means he’s a sneaky snake and worst everyone loved him that’s why I didn’t say anything day one I knew this and It worked and I was fired!

it was a temp job so no big deal but I learned how to deal with co-workers like this are out there and look out and management I knew would be no help but I did tell them but they cared less just like I thought how do you take reports on my training from the person who is training me is not training me so if I don’t know how to do something I get blamed for it?! wtf 2+2 is=4 so why don’t they get that and blame me not him! bs

With that said as nurses let us pull each other up we have enough to deal with that can make us feel we can be at our lowest we don’t need a coworker to speed up the process let’s do better

Don’t dwell on it, especially on people not worthy of your heart or mind. Resiliency is key. Also justifying your reason doesn’t take away from the point that other people might not ever understand your reasons for your actions, especially if they don’t understand why you did it in the first place.

awsome thanks for the advise

I can’t get the questions when I click the button ”start quiz”. What shall I do?

Hi, You need to enable javascript on your browser.

The review was very useful to me. As a student of pharmacy technician, I kindly need more of you.

Question #9 on Part 3 is not correct. I keep getting 1.0281 as the answer

Hi LS, the question also asks to “Record your answer using one decimal place.” so 1.0281 will be 1.1 mL.

1.0281 does not round to 1.1. the second decimal (2) is below 5. It would not round the 1.0 to 1.1. It would stay 1.0 if rounded to the first decimal place

The answer is correct. It’s easy if you set it up like order/on hand then multiply it by the mL.

Desired (D) = 223,500 units Vehicle (V) = 2.5 mL Amount on hand (H) = 500,000 units

Amount to administer (only rounding final answer) = D x V / H = 1.1 mL

Question 19 has be ripping out my hair and maybe someone can explain it to me further. The question states: First, you need to convert 100 mcg/min to mg by moving the decimal point three digits to the left – alternatively, you can divide 100 mcg with 1000 – to get 0.1 mg/min. Why am I dividing by 1000? I thought if we were trying to get a smaller unit of measure to a larger unit of measure we multiply and if we were trying to get a larger unit of measure to a smaller unit we divide. Well MCG if small the MG… wouldn’t we multiply then??

Use unit cancellation method it is much more easier. I got the right answer on my first try. :)

Thanks very much for sharing with us! May the Almighty God bless and protect you in all your undertakings.

I got 95% (1 mistake) which I only forgot to round off. Very nice!

This is very helpful. I get to follow solutions in here. Thank you so much! More power!

Very helpful practice questions.

Was helpful Got only one question but though I haven’t entered school yet but I think I need to learn more on mathematics

We have mcg/min, and we need to get to mL/hour.

First, let’s convert from mcg to mg: 100 mcg/min x 1 mg/1000 mcg = 0.1 mg/min

Next, let’s convert from min to hr: 0.1 mg/min x 60 min/hr = 6 mg/hr

Finally, let’s convert from mg to mL: 6 mg/hr x 500 mL/75 mg = 40 mL/hr

Hope this helps!

The Drug Dosage Calculation Practice Quiz, Question 14: The stated order is for 20mg over an hour. The answer provided and the rationale for the answer reflect a 2mg order.

If possible please correct the answer or the order. I spent some time trying to figure out where I was going wrong. – James

question 14. I’m confused where 2mg came from whilst order stated furosemide (Lasix) 20 mg

2mg/min x 250ml/400mg x 60/hr= 75 ml/hr

Sorry about that, it should be 2mg not 20mg. Item fixed.

please help how to solve 1tabletx0.25\0.125

Awesome ! May ALMIGHTY GOD bless you !

These practice questions help me so much, thank you!

Is question # 24 the right answer?

Question: A health care provider orders diphenhydramine hcl (Benadryl) 180 mg/m2/day to a 12 year old child. The child’s weight is 93 pounds and is 5 feet 2 inches tall. The medication label shows the normal adult dose of 25 mg t.i.d. How many mg of benadryl will the child receive at each dose?

Answer: 19.26 mg/day

If you are wanting to find how much per DOSE, should you divide by 3 doses (t.i.d)?

Are there classes for Nclex

Table of Contents for Dosage Calculations Tutorial

Dosage Question Steps
Unit Conversion
Mass for Mass
Mass/Liquid For Liquid
IV Terms and Abbreviations
Amount in IV Fluid
Volume/Time - IV mL Rate
Volume/Time - IV Drop Rate
Fluid Maintenance Requirement
Dosage By Weight
Mass/Time - IV mL Rate
Practice Questions

Dosage Calculation for Nursing Students: 15 Resources to Help You Practice

Dosage calculations are one of the most critical skills a nursing student learns as it lowers the risk of toxicity and improves patient outcomes.
Not all nursing students have the same learning style, so it’s crucial to teach students in a manner that makes it easiest for them to learn.
These 15 resources incorporate books, online tutorials, videos, audio, worksheets, and quizzes so nursing students can find the best resource.

Are you intimidated by making dosage calculations? These are critical to a patient’s health and recovery. Calculating dosages is likely one of the most important skills you’ll gain as a nursing student and is necessary to ensure medications are administered correctly.

Nursing students need education and practice to learn how to calculate intravenous (IV) drip rates (when the IV pump is broken) and draw up the correct amount of fluid in a syringe. Explore the importance of calculating the right dosage and 15 easily accessible resources that will help you become an expert at this crucial skill.

The Importance of Dosage Calculations in Nursing

Medication dosages are calculated on many factors, including the weight of the patient, diagnosis, age, comorbidities, and risk factors. Manufacturers cannot package drugs for every possible combination. The goal is for the patient to receive the correct medication dosage. Nurses may need to calculate the amount based on the concentration of the medication and the patient’s weight.

Incorrect dosing can lead to adverse events, toxicity, and ineffective treatment. While technology and medication administration software can help, nurses must have a strong understanding and ability to ensure the technology is used correctly.

15 Dosage Calculation Resources for Nursing Students

Calculating the correct dosage takes practice. Finding the right resources for dosage calculations for nursing students can be challenging. You want a tool that explains the reasoning behind the calculations, so you understand and learn the technique. You also want a tool that gives you feedback on the answers.

We scoured the available resources to compile a list that works with various learning styles. Some people like reading books and using worksheets; others may prefer online tools or watching video lectures.

Whatever your preferred learning style, there’s a dosage calculation resource on this list for you.

Online Tools

Dosage help, rmit university, dosage calculation workbook, practice quizzes, registered nurse rn, lincoln memorial university, southwest tech math/science resources, online videos, lecturio nursing, udemy dosage calculations mastery, clinical calculations, calculation of drug dosages, you might be interested in.

Time Management Tips for Nursing Students

Many nursing students have out-of-control schedules and a vanishing social life. These time management strategies can help you achieve your goals and still get some sleep.

15 Common Courses In RN Programs

What classes do you take in nursing school and how can you prepare for them? Learn about common classes required to become an RN and how to succeed.

NCLEX-RN Pass Rates Fall to Lowest Point in the Past Decade

The 2022 NCLEX-RN pass rates are at their lowest point in 10 years. Explore the reasons and how to prepare students or yourself for the exam.

For Students
For Instructors

Whether on a computer or tablet, Calculating Dosages Online makes learning and applying dosage calculations—even reviewing basic math principles—fun and easy! More than 2,550 questions with rationales cover exactly what you need to know for class, clinical, the NCLEX®, and professional practice

SELF-PACED LEARNING

Complete each module at your own pace.

Questions at the end of each module, along with quizzes at the end of each section, reinforce what you've learned and what you haven't.

Safety-focused videos demonstrate exactly how to administer medications properly. Full-color photographs and detailed instructions show you how to read medication labels and analyze syringes, needles, and other equipment.

Two tests for each module and a comprehensive review assess your competency and identify areas for further study.

Practice exercises and review questions help you apply what you're learning, and an Online Workbook in PDF format (with answers) builds your skills with additional practice.

EXISTING STUDENT / ADOPTING FACULTY LOGIN

New student join, forgot password.

An official website of the United States government

The .gov means it's official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you're on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Publications
Account settings
Browse Titles

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2024 Jan-.

StatPearls [Internet].

Dose calculation ratio and proportion method.

Tammy J. Toney-Butler ; Samar Nicolas ; Lance Wilcox .

Affiliations

Last Update: June 20, 2023 .

Introduction

Three primary methods for calculation of medication dosages exist; Dimensional Analysis, Ratio Proportion, and Formula or Desired Over Have Method. We are going to explore the Ratio-Proportion Method, one of these three methods, in more detail.

Ratio-Proportion Method allows us the ability to compare numbers, units of measurement, or values. [1] [2] [3]

Clinicians must define a ratio and proportion. Ratios, often expressed in fraction format, are mathematical works of art designed in relationship patterns that explore comparisons between units, words, numbers. As in any relationship, key players forge a bond to make the association stronger or manageable. Proportions are those key players formed by the equality of ratios. A complicated relationship simplified by utilization and strategic placement of key players of like units or volumes. Ratios and proportions expressed as fractions, canceled out by cross multiplication or division, provide for ease in problem-solving using this method of drug calculation.

Numerators (top) numbers or denominators (bottom) numbers multiplied and divided after the same units are canceled out. Some equations or formulas get expressed with a colon (:) or backslash (/) to indicate division and its subsequent deployment in this problem-solving technique.

For ease of calculation, a person should place the numerator of the fraction to the left of the colon or slash. In completion of this relationship, the denominator gets put on the right of the slash or colon. Unknown amounts, unknown quantities, or unknown desired amounts are depicted as an (x) in the equation and solved. The symbol (x) placement is to the left of the equation, making cross multiplication and division for (x) a simple undertaking. Keeping in mind the fundamental principle regarding the same units of measurement, numbers or units on the top and bottom of a fraction possess the ability to cancel each other out. [4] [5]

It is common for healthcare workers to use a calculator to calculate drug dosages. Calculators may be useful to decrease medication errors related to a calculation issue but are not helpful in the recognition of a conceptual error (Savage, 2015).

One study by Boyle and Eastwood found there were (40%) conceptual errors, (60%) arithmetical errors, and (25%) computational errors amongst study participants in a paper-based drug calculation questionnaire of twenty paramedics where no calculator was allowed (Boyle & Eastwood, 2018).

This study highlighted the need for ongoing education modules to improve basic arithmetic skills, for example, proper use of formulas, ability to construct a mathematical equation, and ability to perform long division without the use of a calculator. Calculators are not always available in a prehospital environment, and cell phone coverage with the use of apps may be spotty at best. The study concluded that a basic knowledge of performing manual drug calculations be a part of training modules by educators (Boyle & Eastwood, 2018).

Using 1 of the 3 methods of drug calculation as discussed above will ease the performance of manual drug calculations; Ratio and Proportion, Desired Over Have or Formula, and Dimensional Analysis. Whichever method is employed by the healthcare provider, a second method may be of use as a check for the first method. A second check further decreases the chance of a medication error related to an incorrect dosage calculation.

Technique or Treatment

Ratio and Proportion Method

The Ratio and Proportion Method has been around for years and is one of the oldest methods utilized in drug calculations. Addition principles are a problem-solving technique that has no bearing on this relationship; only multiplication and division are used to navigate our way through a ratio and proportion problem, not adding. An example listed below will help us provide a better explanation using a fraction or a colon format:

A provider orders lorazepam 4 mg intravenous (IV) push for a CIWA score of 25. On hand, the clinicians has 2 mg/mL vials. How many milliliters are required to carry out the ordered dose?

Have on hand/Quantity you have = Desired Amount/x
2 mg/1 mL = 4 mg/x

In colon format, you would use H:V::D:X and multiply means DV and Extremes HX.

Hx = DV, x = DV/H, 2:1::4:x, 2x = (4)(1), x = 4/2, x = 2 mL

Desired Over Have or Formula Method

Desired over Have or Formula Method uses a formula or equation to solve for an unknown quantity (x), much like ratio proportion. Drug calculations require the use of conversion factors, such as when converting from pounds to kilograms or liters to milliliters. Simplistic in design, this method allows us to work with various units of measurement, converting factors to find our answer (as cited in Boyer, 2002) [Lindow, 2004]. Useful in checking the accuracy of the other methods of calculation as above mentioned, thus acting as a double or triple check.

A basic formula, solving for x, guides us in the setting up of an equation: D/H x Q = x, or desired dose (amount) = ordered dose amount/amount on hand x quantity.

For example, a provider requests lorazepam 4 mg IV Push for a patient in severe alcohol withdrawal. On hand, the clinician has 2 mg/mL vials. How many milliliters should they draw up in a syringe to deliver the desired dose?

Dose ordered (4 mg) x quantity (1 mL)/have (2 mg) = amount you want to give (2 mL)

Remember, units of measurement must match such as milliliters and milliliters, or you will need to convert to like units of measurement. In the example above, the ordered dose was in mg, and the have dose was in mg; both would cancel out, leaving milliliters (answer called for milliliters), so no further conversion is required.

Dimensional Analysis Method

An order placed by a provider for lorazepam 4 mg IV PUSH for CIWA score of 25 or higher, follow CAGE Protocol for subsequent dosages based on CIWA scoring.

On hand, the supply is with 2 mg/mL vials in the automated dispensing unit.
How many milliliters are needed to arrive at the ordered dose?
The desired dose gets placed over 1 remember, (x mL) = 4 mg/1 x 1 mL/2 mg x (4)(1)/2 x 4/2 x 2/1 = 2 mL, the clinician kept multiplying/dividing until they got the desired amount, 2 mL in this problem example
Notice the fraction was set up with mg and mg strategically placed so like units could cancel each other out, making the equation easier to solve for the unit you desired or milliliters.

Zeros can be canceled out in the same way as like units. Look at the example listed below for clarification:

1000/500 x 10/5 = 2, the 2 zeros in 1000 and 2 zeros in 500 can be crossed out since like units in numerator and denominator, leaving 10/5, a much easier fraction to solve, and the answer makes sense.

Having addressed zeros, the following is a look at 1.

If you multiply a number by a 1, then the number is unchanged.
In contrast, if you multiply a number by zero, the number becomes zero.
Examples listed below are as follows: 18 x 0 = 0 or 20 x 1 = 20.
Clinical Significance

Medication errors can be detrimental and costly to patients. Drug calculation and basic mathematical skills play a role in the safe administration of medications. Pediatric populations are especially vulnerable to medication errors due to the need to calculate dosages incorporating many factors; height, weight, body surface area, and growth and development level. The higher the complexity of the math, the increased the risk potential for dose calculation errors.

According to a study of intensive care nurses (ICU) in 2016, 80% of nurses considered knowledge on drug dosage calculation essential to decrease medication errors during the preparation of intravenous drugs.

One study by a group of oncology nurses in 3 Swiss hospitals published in 2018 discusses the process of double-checking and its limitations in the current healthcare environment, specifically, increased nurse workload and time constraints, distracting environments, and lack of resources. The study concluded that oncology nurses strongly believed in the effectiveness of double-checking medication despite reporting limitations of the procedure in clinical practice. [6]

Enhancing Healthcare Team Outcomes

High-risk medications such as heparin and insulin often require a second check on dosage amounts by more than one provider before administration of the drug. Follow institutional policies and recommendations on the double-checking of dose calculations by another licensed provider.

Review Questions
Access free multiple choice questions on this topic.
Comment on this article.

Disclosure: Tammy Toney-Butler declares no relevant financial relationships with ineligible companies.

Disclosure: Samar Nicolas declares no relevant financial relationships with ineligible companies.

Disclosure: Lance Wilcox declares no relevant financial relationships with ineligible companies.

This book is distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0) ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ), which permits others to distribute the work, provided that the article is not altered or used commercially. You are not required to obtain permission to distribute this article, provided that you credit the author and journal.

Cite this Page Toney-Butler TJ, Nicolas S, Wilcox L. Dose Calculation Ratio and Proportion Method. [Updated 2023 Jun 20]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2024 Jan-.

In this Page

Bulk download.

Bulk download StatPearls data from FTP

Related information

PMC PubMed Central citations
PubMed Links to PubMed

Recent Activity

Dose Calculation Ratio and Proportion Method - StatPearls Dose Calculation Ratio and Proportion Method - StatPearls

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

Connect with NLM

National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894

Web Policies FOIA HHS Vulnerability Disclosure

Help Accessibility Careers

DosageMath.com offers free resources to faculty and students to support the learning of medication dosage calculations.

Rather than focusing on memorizing formulas, the resources on DosageMath.com place an emphasis on understanding the "why," developing sound quantitative reasoning, and building confidence in completing dosage calculation tasks.

Please use the links below to access resources designed for faculty and students.

Instructor Resources

V irtual tools to measure and "prepare" dosage for administration

Oral dosage cup

Syringes and syringe cart

IV drip chamber

Images and animations for creating dosage tasks

Links to Desmos Classroom activities

Student Resources (Coming Soon)

This section will be updated soon.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License .

Fastest Nurse Insight Engine

MEDICAL ASSISSTANT
Abdominal Key
Anesthesia Key
Basicmedical Key
Otolaryngology & Ophthalmology
Musculoskeletal Key
Obstetric, Gynecology and Pediatric
Oncology & Hematology
Plastic Surgery & Dermatology
Clinical Dentistry
Radiology Key
Thoracic Key
Veterinary Medicine
Gold Membership

Medications and Calculations

Chapter 14 Medications and Calculations Overview This chapter on medications and calculations is subdivided into six sections: (A) systems of measurement with conversion, (B) methods for calculation, (C) calculations of oral dosages, (D) calculations of injectable dosages, (E) calculations of intravenous (IV) fluids, and (F) pediatric drug calculations. The nurse may proceed independently through Sections A to F to practice and master calculation of drug dosages during the fundamental nursing or pharmacology course. This chapter also serves as a review of drug calculation for nurses in practice settings. Numerous drug labels are used in the drug calculation problems to familiarize the nurse with important information on a drug label. This information is then used in correctly calculating the drug dose. Six calculation methods are explained. Four are general methods: (1) basic formula, (2) ratio and proportion, (3) fractional equation, and (4) dimensional analysis. The nurse should select one of these general methods for the calculation of drug dosages. The other two methods are used to individualize drug dosing by body weight and body surface area. Each calculation method has a color-coded icon that identifies the method used in the chapters. The drug calculation chart in Table 14A-3 may be used in the clinical setting. The nurse might find it helpful to review Chapter 13 on medication administration. Keeping in mind that the goal is to prepare and administer medications in a safe and correct manner, the following recommendations are offered: • Think. Focus on each step of the problem. This applies to simple and difficult problems. • Read accurately. Pay particular attention to the location of the decimal point and to the operation to be done, such as conversion from one system of measurement to another. • Picture the problem. • Identify an expected range for the answer. • Seek to understand the problem. Do not merely master the mechanics of how to do it. Ask for help when unsure of the calculation. Section 14A: Systems of Measurement with Conversion Outline Key Terms Metric System Conversion within the Metric System Metric Conversion Household System Household Conversion Metric, Apothecary, and Household Equivalents Objectives • Discuss the two systems of measurement. • Convert measurement within the metric system, larger units to smaller units, and smaller units to larger units. • Convert measurements within the household system, larger units to smaller units, and smaller units to larger units. • Convert metric, apothecary, and household measurements among the three systems of measurement as appropriate. See Table 14A-3 . Key Terms apothecary system, p. 150 gram, p. 149 household measurement, p. 149 liter, p. 149 meter, p. 149 metric system, p. 149 minim, p. 150 ounce, p. 149 Two systems of measurement—metric and household—are used to measure drugs and solutions. The metric system, developed in the late eighteenth century, is the internationally accepted system of measure. It is replacing the apothecary system, which dates back to the Middle Ages and had been used in England since the seventeenth century. Household measurement is commonly used in community and home settings in the United States. Metric System The metric system is a decimal system based on the power of 10. The basic units of measure are gram (g, gm, G, Gm) for weight; liter (l, L) for volume; and meter (m, M) for linear measurement, or length. Prefixes indicate the size of the units in multiples of 10. Table 14A-1 gives the metric units of measurement in weight (gram), volume (liter), and length (meter) in larger and smaller units that are commonly used. TABLE 14A-1 METRIC UNITS OF MEASUREMENT UNIT NAMES AND ABBREVIATIONS MEASUREMENTS Gram (weight) 1 kilogram (kg, Kg) 1000 g 1 gram (g, gm, G, Gm) 1 g 1 milligram (mg) 0.001 g 1 microgram (mcg) 0.000001 g 1 nanogram (ng) 0.000000001 g Liter (volume) 1 kiloliter (kl, KL) 1000 L (l) 1 liter (L, l) 1 L (l) 1 milliliter (mL) 0.001 L (l) Meter (length) 1 kilometer (km) 1000 m 1 meter (m, M) 1 m 1 centimeter (cm) 0.01 m 1 millimeter (mm) 0.001 m NOTE: 1 mL (milliliter) = 1 cc (cubic centimeter). Values are the same in drug and fluid therapy. 1 mg (milligram) = 1000 mcg (micrograms). Kilo is the prefix used for larger units (e.g., kilometer), and milli, centi, micro, and nano are the prefixes used for smaller units (e.g., millimeter). The prefix stands for a specific degree of magnitude; for instance, kilo stands for thousands, milli for one thousandth, centi for one hundredth, and so on. Because the difference between degrees of magnitude is always a multiple of 10, converting from one magnitude to another is relatively easy. Conversion within the Metric System The metric units most frequently used in drug notation are the following: 1 g = 1000 mg 1 L = 1000 mL 1 mg = 1000 mcg To be able to convert a quantity, one of the values must be known, such as gram or milligrams, liters or milliliters, and milligrams or micrograms. Gram, liter, and meter are larger units; milligram, milliliter, and millimeter are smaller units. Metric Conversion A. When converting larger units to smaller units in the metric system, move the decimal point one space to the right for each degree of magnitude change. Note: This does not apply to micro and nano units. Example Change 1 gram to milligrams. Grams are three degrees of magnitude greater than milligrams (see Table 14A-1 ). Move the decimal point three spaces to the right. B. When converting smaller units to larger units in the metric system, move the decimal point one space to the left for each degree of magnitude of change. Example Change 1000 milligrams to grams. Milligrams are three degrees of magnitude smaller than grams. Move the decimal point three spaces to the left. REMEMBER: When changing larger units to smaller units, move the decimal point to the right, and when changing smaller units to larger units, move the decimal point to the left. Practice Problem 1 Metric Conversion Larger to Smaller Units Smaller to Larger Units 1. Change 2 g to mg 4. Change 1500 mg to g 2. Change 0.5 ( ) g to mg 5. Change 3 g to kg 3. Change 2.5 L to mL 6. Change 500 mL to L Household System Household measurement is not as accurate as the metric system because of the lack of standardization of spoons, cups, and glasses. The measurements are approximate. A teaspoon (t) is considered to be equivalent to 5 mL according to the official United States Pharmacopeia. Milliliters (mL) are the same as cubic centimeters (cc) in value. Three teaspoons equals 1 tablespoon (T). Ounces (oz) are fluid ounces in the household measurement system; the word fluid in front of ounce is usually not used. One milliliter of water fills a cubic centimeter exactly. Table 14A-2 gives the household equivalents in fluid volume. The measurements with asterisks are frequently used in drug therapy and should be remembered. TABLE 14A-2 HOUSEHOLD EQUIVALENTS IN FLUID VOLUME 1 MEASURING CUP = 8 OUNCES (oz) 1 medium-size glass (tumbler size) = 8 ounces (oz) 1 coffee cup (c) = 6 ounces (oz) (varies with cup size) 1 ounce (oz) = 2 tablespoons (T) 1 tablespoon (T) = 3 teaspoons (t) 1 teaspoon (t) = 60 drops (gtt) * 1 drop (gt) * = 1 minim (min, or m) * Varies with viscosity of liquid and dropper opening. Household Conversion A. When converting larger units to smaller units within the household system, multiply the requested number by the basic equivalent value. Example Change 2 glasses of water to ounces. The equivalent value is 1 medium-sized glass = 8 oz (fl oz) 2 glasses × 8 oz = 16 oz (fl oz) With discharge teaching for a patient who requires liquid medication(s) at home, the nurse may find it necessary to convert metric measurements to household measurements. Practice Problem 2 Household Conversion REMEMBER: To change larger units to smaller units, multiply the requested number of units by the basic equivalent value. To change smaller units to larger units, divide the requested number of units by the basic equivalent value. Refer to Table 14A-2 as needed. Larger to Smaller Units Smaller to Larger Units 1. Change 3 oz to T ______ 4. Change 3 T to oz ______ 2. Change 5 T to t ______ 5. Change 16 oz to a measuring cup ______ 3. Change 3 coffee cups to oz ______ 6. Change 12 t to T ______ Metric, Apothecary, and Household Equivalents Although the apothecary system is no longer used, Table 14A-3 is included to show metric and apothecary equivalents by weight, and metric, apothecary, and household equivalents by volume. Practice Problem 3 Summary: Metric and Household Measurements Metric System: Refer to Table 14A-1 as needed. 1. 2 g = _________ mg 5. 500 mg = _________ g 2. 1.2 kg =________ g 6. 10,000 mcg = _____ mg 3. 5 mg = _______ mcg 7. 2400 mg = ________ g 4. 2.5 L = _______ mL 8. 1500 mL = ________ L Household System: Refer to Table 14A-2 as needed. 1. 5 glasses = _____ oz 4. 4 oz = ___________ T 2. 3 T = __________ t 5. 15 t = ___________ T 3. 2 c = _________ oz 6. 5 T = ___________ oz TABLE 14A-3 APPROXIMATE METRIC, APOTHECARY, AND HOUSEHOLD EQUIVALENTS METRIC SYSTEM APOTHECARY SYSTEM HOUSEHOLD SYSTEM Weight 1 kg 1000 g 2.2 lb 2.2 lb * 1 g 1000 mg 15 (16) gr 0.5 g 500 mg gr 0.3 g 300 (325) mg 5 gr 0.1 g 100 mg gr * 0.06 g 60 (65) mg 1 gr 0.03 g 30 (32) mg gr 0.01 g 10 mg gr 0.6 mg gr 0.4 mg gr 0.3 mg gr Volume 1 L; 1000 mL 1 qt; 32 oz (fl oz) 0.5 L; 500 mL 1 pt; 16 oz (fl oz) 0.24 L; 240 mL 8 fl oz 1 glass 0.18 L; 180 mL 6 fl oz 1 c * 30 mL 1 fl oz; 8 fl dr 2 T; 6 t 15 mL fl oz; 4 fl dr 1 T; 3 t † 5 mL 1 t 4 mL 1 fl oz; 60 (min) 1 t 1 mL 15 (16) 15-16 gtt Height/distance 2.54 cm 1 in 1 in 25.4 mm 1 in 1 in fl dr , Fluid dram; fl oz , fluid ounce; minim; cm , centimeter; g , gram; gr , grain; gtt , drops, in , inch; kg , kilogram; L , liter; lb , pound; mg , milligram; mL , milliliter; mm , millimeter; pt , pint; qt , quart; T , tablespoon; t , teaspoon; c , coffee cup. * Equivalents commonly used for computing conversion problems by ratio. † 5 mL = 1 t (teaspoon); official United States Pharmacopeia measurement. Answers to Practice Problems 1 Metric Conversion 1. 2.0 g = or 2.0 g = 2000 mg The gram is three degrees of magnitude greater than the milligram, so the decimal point is moved three spaces to the right. 2. 0.5 g = or 0.5 g = 500 mg The gram is three degrees of magnitude greater than the milligram, so the decimal point is moved three spaces to the right. 3. 2.5 L = or 2.5 L = 2500 mL The liter is three degrees of magnitude greater than the milliliter, so the decimal point is moved three spaces to the right. 4. 1500 mg = or 1500 mg = 1.5 g The milligram is three degrees of magnitude smaller (less) than the gram, so the decimal point is moved three spaces to the left. 5. 3 g = or 3 g = .003 mg The gram is three degrees of magnitude smaller than the kilogram, so the decimal point is moved three spaces to the left. 6. 500 mL = or 500 mL = 0.5 L The milliliter is three degrees of magnitude smaller than the liter, so the decimal point is moved three spaces to the left. 2 Household Conversion 1. 3 oz = 6  T; the equivalent value is 3 oz × 2 = 6 T 2. 5 T = 15  t; the equivalent value is 1 T = 3 t 3. 3 c = 18  oz; the equivalent value is 1 c = 6 oz 4. 3 T = oz; the equivalent value is 3 T ÷ 2 or oz 5. 16 oz = 2  c; the equivalent value is 1 measuring cup = 8 oz 6. 12 t = 4  T; the equivalent value is 1 T = 3 t 3 Summary Metric and Household Measurements Metric Conversion 1. 2000 mg (1 g = 1000 mg) 2 × 1000 mg = 2000 mg or (three spaces to the right) 2. 1200 g 3. 5000 mcg (1 mg = 1000 mcg) 4. 2500 mL 5. 0.5 g (1000 mg = 1 g) 500 ÷ 1000 = 0.5 or = 0.5 g (three spaces to the left) 6. 10 mg 7. 2.4 g 8. 1.5 L Household Conversion 1. 40 oz 2. 9 t 3. 12 oz 4. 8 T 5. 5 T 6. oz Section 14B: Methods for Calculation Objectives • Select a formula—the basic formula, the ratio-and-proportion method, fractional equation, or dimensional analysis—for calculating drug dosages. • Convert all measures to the same system and same unit of measure within the system before calculating drug dosage. • Calculate drug dosage using one of the general formulas. • Calculate drug dosage according to body weight and body surface area. • Discuss meanings for abbreviations used in drug therapy. Outline Interpreting Oral and Injectable Drug Labels Method 1: Basic Formula (BF) Method 2: Ratio and Proportion (RP) Method 3: Fractional Equation (FE) Method 4: Dimensional Analysis (DA) Method 5: Body Weight (BW) Method 6: Body Surface Area (BSA) BSA with the Square Root Key Terms basic formula, p. 153 body surface area (BSA), p. 157 body weight (BW), p. 156 dimensional analysis (DA), p. 155 drug label, p. 152 fractional equation, p. 154 ratio and proportion, p. 153 The four general methods for the calculation of drug doses are (1) basic formula, (2) ratio and proportion, (3) fractional equation, (4) and dimensional analysis. These methods are used to calculate oral and injectable drug doses. The nurse should select one of the methods to calculate drug doses and use that method consistently. For drugs that require individualized dosing, calculation by body weight (BW) or by body surface area (BSA) may be necessary. In the past, these two methods, (5) and (6), have been used for the calculation of pediatric dosages and for drugs used in the treatment of cancer (antineoplastic drugs). BW and BSA methods of calculation are especially useful for individuals whose BW is low, who are obese, or who are older adults. Before calculating drug doses, all units of measure must be converted to a single system (see Section 14A ). It is most helpful to convert to the system used on the drug label . If the drug is ordered in grams (g, G) and the drug label gives the dose in milligrams (mg), then convert grams to milligrams (the measurement on the drug label) and proceed with the drug calculation. Nursing programs prefer to use only the metric system for drug calculations. Interpreting Oral and Injectable Drug Labels Pharmaceutic companies usually label their drugs with the brand name of the drug in large letters and the generic name in smaller letters. The dose per tablet, capsule, or liquid (for oral and injectable doses) is printed on the drug label. Two examples of drug labels are given below, the first for an oral drug and the second for an injectable drug. Example 1 Oral Drug Tagamet is the brand (trade) name, cimetidine is the generic name, and the dose is 200 mg/tablet. Example 2 Injectable Drug Compazine is the brand (trade) name, prochlorperazine is the generic name, and the dose is 5 mg/mL injectable. Method 1 Basic Formula (BF) The basic formula is easy to recall and is most frequently used in calculating drug dosages. The basic formula is the following: D H × V = A where D is the desired dose (i.e., drug dose ordered by the health care provider), H is the on-hand dose (i.e., drug dose on label of container [bottle, vial]), V is the vehicle (i.e., drug form in which the drug comes [tablet, capsule, liquid]), and A is the amount calculated to be given to the patient. Examples 1. Order: cefaclor (Ceclor) 0.5 g PO b.i.d. Available: a. The unit of measure that is ordered (grams) and the unit on the bottle (milligrams) are from the same system of measurement—the metric system. Conversion to the same unit is necessary to work the problem. Because the bottle is in milligrams, convert grams to milligrams. To convert grams (large value) to milligrams (smaller value), move the decimal point three spaces to the right (see Section 14A: Systems of Measurement with Conversion ). b. 2. Order: codeine 60 mg PO STAT Available: a. Method 2 Ratio and Proportion (RP) The ratio and proportion method is the oldest method currently used in the calculation of drug dosages. The formula is as follows: where H is the drug on hand (available), V is the vehicle or drug form (tablet, capsule, liquid), D is the desired dose (as ordered), x is the unknown amount to give, and :: stands for “as” or “equal to.” Multiply the means and the extremes. Solve for x; x is the divisor. Examples 1. Order: amoxicillin (Amoxil) 100 mg PO q.i.d. Available: a. Conversion is not needed because both are expressed in the same unit of measure. b. Answer: amoxicillin 100 mg = 2 mL 2. Order: aspirin/ASA 600 (650) mg q4h PRN Available: aspirin 325 mg/tablet Answer: aspirin 650 mg = 2 tablets Method 3 Fractional Equation (FE) The fractional equation method is similar to ratio and proportion except it is written as a fraction. Cross-multiply and solve for x. Examples Order: ciprofloxacin (Cipro) 500 mg PO q12h Available: How many tablet(s) should the patient receive per dose? Answer: Cross-multiply and solve for x. 250 x = 500 x = 2 tablets of Cipro per dose Method 4 Dimensional Analysis (DA) Dimensional analysis (DA) is a calculation method known as units and conversions. The advantage of DA is that it decreases the number of steps required to calculate a drug dosage. It is set up as one equation. Examples 1. Identify the unit/form (tablet, capsule, mL) of the drug to be calculated. If the drug comes in tablet, then tablet = (equal sign). 2. The known dose and unit/form from the drug label follows the equal sign. Order: amoxicillin 500 mg On the drug label: 250 mg per capsule capsule = 1 cap 250 mg 3. The mg (250 mg) is the denominator, and it must match the next numerator, which is 500 mg (desired dose or order). The NEXT denominator would be 1 (one) or blank. 4. Cancel out the mg, 250 and 500. What remains is the capsule and 2. Answer: 2 capsules When conversion is needed between milligrams (drug label) and grams (order), then a conversion factor is needed, which appears between the drug dose on hand (drug label) and the desired dose (order). Metric Equivalent 1 g = 1000 mg 1 mg = 1000 mcg

Stay updated, free articles. Join our Telegram channel

Comments are closed for this page.

Full access? Get Clinical Tree

A medication dosage simulation strategy to improve patient safety

Simulation boosts skills and encourages active learning. .

Takeaways:

Traditional medication administration dosage testing can be time-consuming, costly, and lack elements of active learning.
Simulation in nurse education provides authentic clinical situations in a safe environment.
A medication dosage simulation project has shown positive outcomes.

Evidence shows that simulation education provides authentic clinical situations in a safe environment and is a link to safe practice. The National Council of State Boards of Nursing (NCSBN) conducted a national multisite study with 666 students to examine simulation use in prelicensure nursing programs. The study was conducted in three phases: simulation usage, student outcomes based on percentage of simulation vs . clinical time, and assessment of graduates as new RNs. The study supported the effectiveness of simulation to gauge competency of prelicensure nurses as they moved into the RN role.

Because simulation has been shown to be an effective teaching tool, we decided to use it to improve patient safety related to medication dosages. We followed simulation guidelines from the International Nursing Association for Clinical Simulation (INA CS L) and the INA CS L Standards of Best Practice: Simulation SM and used the Quality and Safety Education for Nursing competencies to guide our work . Here is an overview of the project and its outcomes .

Case study

Nursing educators met to discuss the ineffectiveness of pediatric and obstetric medication dosage administration courses. The team discussed their frustration with the requirement to repeatedly administer exams with dosage calculations until students achieve a 100% so they can participate in the clinical experience. In addition, s tudents evaluated the experience poorly and described the process as a high – anxiety event.

The team of educators decided to implement medication dosage administration simulation stations with competency assessments . Each simulation station is attended by a faculty member familiar with the content , and the student’s calcu l ations are verified before he or she moves to the next station. After completing all of the stations, each student take s a medication dosage calculation exam and participate s in a reflection exercise. (See Reflecting.)

The reflection questions were designed to help students consider the importance of medication calculation and administration. The first two questions are used for students who did not achieve 100% on the exam.

If you made an error, review the calculation and consider how your error occurred.
List three steps you must take to relearn and recalculate.

Now that you’ve completed questions 1 and 2, return to the station for remediation. When you have accomplished 100% accurate remediation, complete a second medication dosage calculation exam. Retest only on the area(s) where you didn’t achieve 100%.

The second two questions are used for those who achieved 100% on the exam.

What reaction would you expect to have if you made a medication error during your clinical practice and why?
List three steps you must take to correct any medication error in clinical practice.

If the student doesn’t achieve 100% on the exam , he or she return s to the appropriate station or stations to remediate the concept and practice dosage calculations. The student is then retest ed on the same day in the area where an error was made. If a student doesn’t achieve 100% after remediation, a second round of remediation and retesting occur s .

Pre – implementation planning and challenges

Before implementing the simulation, several planning steps were completed , including notifying the program directors and course faculty of the change. Most accepted the new format, but one faculty member was concerned that the simulations and remediation weren’t rigorous enough. She felt that the repetitive math testing should continue with no simulations or remediation. When the process was more fully explained and the faculty member understood that all students would still need to achieve 100% on the exam , she agreed to the new format. She was invited to join the team and participate in the stations during the lab.

To ensure all students could attend the simulation lab, some scheduling changes were required. One team member worked with a faculty member to switch the time of a class to incorporate the lab into the student schedule before the beginning of the clinical experience. Another team member divided the students into groups of eight and plotted the flow of each 90-minute lab session with four practice stations and one testing station.

Additional rooms and space were found and reserved , and time was established in the course schedule and in the simulation center for the new program . In addition, more supplies were purchased for the lab , faculty were identified to attend the simulation stations, a plan was created for setting up and dismantling the stations , student guidelines were developed , and a post activity evaluation and student reflection were created .

To prepare students for the simulation, they were provided with readings, an online module with narration, and practice problems. Four different tests were developed to prevent student s from sharing information with the next group.

Planning for the simulation lab began in September 2017, organization occurred in October, and supplies were ordered in December. In January 2018, the new strategy was implemented with the BSN students .

Objectives and o utcomes

The objectives for medication dosage administration lab already existed for the pediatric/obstetrics course, but they were revised to include the new simulation format .

Objective 1. S tudents will calculate accurate medication dosages and prepare medications safely by the end of the simulation lab.

Measurable goal : All students enrolled in the pediatric/obstetrics course will successfully complete the revised medication dosage administration simulation lab .
Results : A total of 99 students successfully completed the medication dosage administration simulation lab. Some students had to remediate and complete the second math test on the same day as the lab.

Objective 2 . S tudents will demonstrate safe and accurate psychomotor skills by the end of the medication dosage administration simulation lab.

Measurable goal : All students will perform math calculations and medication dosage administration at each lab station and demonstrate competency.
Results : All students practiced psychomotor and safety skills related to medication dosage preparation, medication administration , and math calculation at each lab station. A faculty member at each station checked students’ competencies before the y proceeded to the next station.

Objective 3 . S tudents will complete the medication administration dosage exam with 100% accuracy.

Measurable g oal : All students will complete the medication dosage exam with 100% accuracy on the first or second attempt on the same day as the scheduled lab.
Results : On the first attempt at the exam after the simulation activity, 45% of students achieved 100% accuracy. The remaining students passed the medication dosage exam on the second attempt after remediation at the appropriate station. All students started their clinical experience on time.

Objective 4 . S tudents will complete a reflection about safe medication administration at the end of the simulation lab.

Measurable goal : All students will complete the reflection.
Results : All students completed a reflection and then were dismissed from the lab. Students commented that the reflection made them review their performance and realize the importance of safe medication administration.

Objective 5 . S tudents will complete an online evaluation one day after the medication dosage administration simulation lab.

Measurable goal : All students will complete the online evaluation of the lab.
Results : The online evaluation was completed by 50% of the students. The data were analyzed and improvements made for the next lab. Faculty also completed an evaluation within 1 week.

Simulation day

T o make the simulation as realistic as possible, equipment and scenarios were created to be similar to a real-life hospital setting. Each skill station was equipped with hand sanitizer, gloves, alcohol wipes, sharps containers, and calculators. A pediatric or obstetric faculty member was at each station to answer questions and monitor and verify accurate skill completion .

Groups of eight students each rotated through the lab at scheduled times . They were given a 90-minute time limit , with 15 minutes for each station (an extra 15 minutes was available if the group needed it ). As students entered the room, they signed in and took a folder with instructions and a math problem for each station. To ensure all students had completed the prelab activities, each sign ed a statement signifying they had viewed the online module and brought completed practice problems to the lab for admission.

At each station ( fluid maintenance and urine output, safe dosing, reconstitution, and drip rate ) , students calculated the correct dose and practiced preparing and administering the medication. The final exam and reflection were administered in a separate, quiet room next to the lab. Students had 15 minutes to complete the exam.

F our math exams were developed by the team and distributed randomly to student s when they entered the final station. The exam s consisted of math problems similar to those completed at each station. If a student scored 100% on the exam, he or she completed the two reflection questions. If a student didn’t score 100%, he or she returned to the simulation area and remediated at the stations related to the exam errors . The student then returned to the testing room to retake the portion s of the exam related to the error s and completed the reflection questions . Upon completion, the student was considered ready to administer medications in the clinical setting .

The new medication dosage administration simulation lab has been completed with two groups of students . Evaluations have been positive , with areas of improvement suggested by students and instructors . (See Simulation evaluation. )

Simulation evaluation

Students and instructors were asked to complete a simulation lab evaluation.

Student results and comments

Was the process of preparing for the day clear? Yes: 91%
Was the process during the day clear? Yes: 91%
Did your knowledge of medication safety and delivery improve? Yes: 95%
How much do you value medication safety? 78% valued it more than before the simulation lab
Was safety day a positive experience? Yes: 87%
Thank you!”
“Very helpful and fun.”
“Tell us up front what the objectives are.”
“Waiting at stations is boring.”
“Conducive to the way we learn.”
“Relaxed environment and patient faculty.”
“Executed really well!”

Instructor comments

Well organized but need to revise practice math problems to better reflect actual dosages.
One station reviewing a patient safety poster can be included in the pre simulation presentation.
Realistic supplies are needed and worked well.
Pre simulation assignment should be better explained.

At the end of the lab, students are ready to administer medications in clinical , and they don’t have to undergo repetitive math testing. The new process is more efficient for students and faculty. The working group of nurse educators continue s to make improvements with the simulation. For example, they’ve arranged for smaller groups and clarif ied the pre simulation instructions.

This learning strategy can be used with prelicensure students, graduate nurses, preceptees , and staff nurses in a variety of specialties. In all cases, patient safety related to medication dosage administration and calculation can be enhanced.

Nadine M. Marchi is a clinical assistant professor at The George Washington University School of Nursing in Ashburn, Virginia. At Case Western Reserve Frances Payne Bolton School of Nursing in Cleveland, Ohio, Elizabeth Zimmermann is an assistant professor, and Connie S. Kelling, Shannon Wong, and Kathleen M. Juniper are nursing instructors.

Selected references

Benner P. Curricular and pedagogical implications for the Carnegie Study, educating nurses: A call for radical transformation. Asian Nurs Res . 2015;9 (1) :1-6.

Cant RP, Cooper SJ. The value of simulation-based learning in pre-licensure nurse education: A state-of-the-art review and meta-analysis. Nurs e Educ Pract . 2017;27:4 5 -6 2 .

Huston CL, Phillips B, Jeffries P, et al. The academic – practice gap: Strategies for an enduring problem. Nurs Forum . 2018;53 (1) :27-34.

Jeffries PR, Dreifuerst KT, Kardong-Edgren S, Hayden J. Faculty development when initiating simulation programs: Lessons learned from the National Simulation Study. J Nurs Reg ul . 2015;5(4):17-23.

Lindell A. Enhancing medication safety teaching through remediation and reflection. QSEN Institute. November 1, 2016 . qsen.org/enhancing-medication-safety-teaching-through-remediation-and-reflection

Miraglia R , Asselin ME. Reflection as an educational strategy in nursing professional development : An integrative review . J Nurses Prof Dev . 2015;31(2):62-7 2 .

1 Comment .

I find this tool to be very useful and should be incorporated into the healthcare setting for nurses education day to keep everyone abreast on how important it is for patient safety when it comes to the delivery of medication. Different nursing departments lose their skills if they are not giving medications routinely (i.e., psych. unit, and the ortho rehab., versus the ICU/CCU, open-heart unit, neuro-intensive care, surgical intensive care). Thank you for the article.

Comments are closed.

NurseLine Newsletter

First Name *
Last Name *
Hidden Referrer

*By submitting your e-mail, you are opting in to receiving information from Healthcom Media and Affiliates. The details, including your email address/mobile number, may be used to keep you informed about future products and services.

Test Your Knowledge

Competent care during pregnancy loss

Knowledge of intravascular determination

Why Mentorship?

Virtual reality: Treating pain and anxiety

What is a healthy nurse?

Broken trust

Diabetes innovations and access to care

Breaking barriers: Nursing education in a wheelchair

Chronic traumatic encephalopathy and traumatic brain injury

Supporting the multi-generational nursing workforce

Vital practitioners

From data to action

Connecting theory and practice

Stevens-Johnson syndrome

Turning the tide

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

View all journals
My Account Login
Explore content
About the journal
Publish with us
Sign up for alerts
Open access
Published: 10 May 2024

MISATO: machine learning dataset of protein–ligand complexes for structure-based drug discovery

Till Siebenmorgen ORCID: orcid.org/0009-0008-5160-8100 1 , 2 na1 ,
Filipe Menezes ORCID: orcid.org/0000-0002-7630-5447 1 , 2 na1 ,
Sabrina Benassou 3 ,
Erinc Merdivan 4 ,
Kieran Didi ORCID: orcid.org/0000-0001-6839-3320 5 ,
André Santos Dias Mourão 1 , 2 ,
Radosław Kitel 6 ,
Pietro Liò 5 ,
Stefan Kesselheim ORCID: orcid.org/0000-0003-0940-5752 3 ,
Marie Piraud 4 ,
Fabian J. Theis ORCID: orcid.org/0000-0002-2419-1943 4 , 7 , 8 ,
Michael Sattler ORCID: orcid.org/0000-0002-1594-0527 1 , 2 &
Grzegorz M. Popowicz ORCID: orcid.org/0000-0003-2818-7498 1 , 2

Nature Computational Science ( 2024 ) Cite this article

7885 Accesses

35 Altmetric

Metrics details

Computational biophysics
Drug discovery
Machine learning

A preprint version of the article is available at bioRxiv.

Large language models have greatly enhanced our ability to understand biology and chemistry, yet robust methods for structure-based drug discovery, quantum chemistry and structural biology are still sparse. Precise biomolecule–ligand interaction datasets are urgently needed for large language models. To address this, we present MISATO, a dataset that combines quantum mechanical properties of small molecules and associated molecular dynamics simulations of ~20,000 experimental protein–ligand complexes with extensive validation of experimental data. Starting from the existing experimental structures, semi-empirical quantum mechanics was used to systematically refine these structures. A large collection of molecular dynamics traces of protein–ligand complexes in explicit water is included, accumulating over 170 μs. We give examples of machine learning (ML) baseline models proving an improvement of accuracy by employing our data. An easy entry point for ML experts is provided to enable the next generation of drug discovery artificial intelligence models.

Accurate structure prediction of biomolecular interactions with AlphaFold 3

Highly accurate protein structure prediction with AlphaFold

De novo generation of multi-target compounds using deep generative chemistry

In recent years, artificial intelligence (AI) predictions have revolutionized many fields of science. In structural biology, AlphaFold2 (ref. 1 ) predicts accurate protein structures from amino-acid sequences only. Its accuracy nears state-of-the-art experimental data. The success of AlphaFold2 is made possible due to a rich database of nearly 200,000 protein structures that have been deposited and are available in the Protein Data Bank (PDB) 2 . These structures were determined over the past decades using X-ray crystallography, nuclear magnetic resonance (NMR) or cryo-electron microscopy. Despite enormous investments, there are still few new drugs approved yearly, with development costs reaching several billion dollars 3 . An ongoing grand challenge is rational, structure-based drug discovery (DD). Compared with protein structure prediction, this task is substantially more difficult.

In the early stages of DD, structure-based methods are popular and efficient approaches. The biomolecule provides the starting point for rational ligand search. Later, it guides optimization to optimally explore the chemical combinatorial space 4 while still ensuring drug-like properties. In silico methods that are in principle able to tackle structure-based DD include semi-empirical quantum mechanical (QM) methods 5 , molecular dynamics (MD) simulations 6 , 7 , docking 8 and coarse-grained simulations 9 , which can also be combined to be more efficient. However, these methods either suffer from generally low precision or are computationally too expensive while still requiring substantial experimental validation. Recent examples show that classical, ball-and-stick atomistic model representations of biomolecular structures might be too inaccurate in certain situations to allow for correct predictions 10 , 11 , 12 , 13 .

The introduction of AI into the process is still at an early stage. AI approaches are, in principle, able to learn the fundamental state variables that describe experimental data 14 . Thus, they are likely to abstract from electronic and force field-based descriptions of the protein–ligand complex. However, so far mostly simple solutions have been proposed that do not incorporate the available protein–ligand data to their full extent, such as scoring protein–ligand Gibbs free energies 15 , 16 , ADME (absorption, distribution, metabolism and excretion) property estimation 17 or prediction of synthetic routes 18 , 19 . Most of these approaches are constructed using one-dimensional SMILES (simplified molecular-input line-entry system) 20 , 21 and only a few attempts have been made to properly tackle three-dimensional (3D) biomolecule–ligand data 22 , 23 , 24 .

Several databases are available that contain raw experimental structures of protein–ligand complexes, usually extracted from the PDB (for example, PDBbind 25 , bindingDB 26 , Binding MOAD 27 , Sperrylite 28 ). Only recently a database of MD-derived traces of protein–ligand structures was reported 29 , 30 . Despite these efforts, so far no AI model has been proposed that convincingly addresses the rational DD challenge in the way that AlphaFold2 answered the protein structure prediction problem 31 , 32 .

In addition to DD, the structure-based AI models are useful for biomolecule structure analysis and quantum chemistry. However, they are severely hindered by several factors: neglecting the conformational flexibility (dynamics and induced fit upon binding); entropic considerations; inaccuracies in the deposited structural data (incorrect atom types due to missing hydrogen atoms, incorrect evaluation of functional group flexibility, inconsistent geometry restraints, fitting errors); chemical complexity (for example, non-obvious protonation states); overly simplified atomic properties; highly complex energy landscapes in molecular recognition by their targets. Attempts to train AI models currently require inferring this missing information implicitly. The limited number of publicly available protein–ligand structures ( ~ 20,000) and lack of thermodynamic data cause this inference to fail. This is preventing structure-based models from producing groundbreaking results 31 , 32 .

Here, we propose a protein–ligand structural database, MISATO (molecular interactions are structurally optimized) that is based on experimental protein–ligand structures. We show that the database helps to better train models across fields related to DD and beyond. This includes quantum chemistry, general structural biology and bioinformatics. We provide quantum-chemical-based structural curation and refinement, including regularization of the ligand geometry. We augment this database with missing dynamic and chemical information, including MD on a timescale allowing the detection of transient and cryptic states for certain systems. The latter are very important for successful drug design 33 . Thus, we supplement experimental data with the maximum number of physical parameters. This eases the burden on AI models to implicitly learn all this information, allowing focus on the main learning task. The MISATO database provides a user-friendly format that can be directly imported into machine learning (ML) codes. We also provide various preprocessing scripts to filter and visualize the dataset. Example AI baseline models are supplied for the calculation of quantum chemical properties (chemical hardness and electron affinity), for binding affinity calculation and for the prediction of protein flexibility or induced-fit features to simplify adoption. The QM, MD and AI baseline models are validated extensively on experimental data. We wish to transform MISATO into an ambitious community project with vast implications for the whole field of DD.

MISATO dataset

The basis for MISATO (Fig. 1 ) is the 19,443 protein–ligand structures from PDBbind 25 . These structures were experimentally determined over the past decades and represent a diverse set of protein–ligand complexes for which experimental affinities are available. In the context of AI for DD it is of utmost importance to train the models on a dataset with the highest possible correctness and consistency, for several reasons. First, the total number of available structures is much lower than typical training sizes of other AI targets. Second, ligand association has a rather complex energy landscape during molecular recognition. Delicate deviations in the protein–ligand structures or atomic parameters can markedly impair binding. In the PDB, incorrect atom assignments and inconsistent geometries are not uncommon. More seriously, hydrogen atoms are highly sensitive to their chemical and molecular environment and are rarely experimentally accessible. All these issues have been systematically addressed in our work and are compiled in our database (Figs. 2 and 3 ).

a , We provide a dataset that combines semi-empirical QM properties of small molecules with MD-simulated dynamics of the entire set of experimental protein–ligand complexes. All common errors in protein and ligand nomenclature, protonation, geometry and so on are fixed. The blue outline of the molecule describes its electronic density. b , An overview of the dataset and the applied protocols for semi-empirical QM and FF (force field) MD simulations including data preparation, preprocessing and AI baseline models.

a , Statistical overview of changes introduced by our optimization protocol. N atoms corresponds to total changes in the atom count when compared with the source database. In most cases atoms were removed ${N}_{{\mathrm{atoms}}}^{{\mathrm{rem}}}$ ); in only 27% of cases was the number of atoms increased, ${N}_{{\mathrm{atoms}}}^{{\mathrm{add}}}$ . Similar considerations apply to protons—light blue; N protons . b , D4 polarizability versus partial charge for all the halogens in the database. The outliers were analyzed to find possible wrong atom assignments. This was the case for the bromine atom in the lower right corner, which in reality is a boron. c , Examples of inconsistent structures: 1WUG contains overly elongated NO bonds; 4MDN contains a nitrogen in angular violation of VSEPR; 5GTR shows a typical problem in the protonation state. d , e , Calculated electronic density for ketamine (4G8H) and tramadol, respectively (dashed green lines). Dashed circles show the sizes of electronic density around selected atoms. The numbers next to these atoms represent partial charge (top) and atomic polarizability (bottom). These are electronic descriptors representing the electronic density around each center. Color and character keys: N, blue, nitrogen; S, yellow, sulfur; O, red, oxygen; C, beige, carbon; H, white, hydrogen; Cl, green, chlorine.

Source data

a – c , Reversible opening and closing of the binding pocket can be captured during the simulations, including cryptic binding sites. a , The structure of 2AM4 is shown after 2 ns (left panel), 6 ns (middle panel) and 10 ns (right panel) simulation time (fluorine in beige). b , c , The opening loop region ( b , structure 2LKK) is visualized for superimposed timesteps (blue diagram, dark hue, 2 ns; black diagram, medium dark hue, 6 ns; red diagram, light hue, 10 ns). The protein pocket opens in structure 8ABP during the simulation ( c ). d , Protein residues at the binding site can undergo large adaptations within the simulations, indicating unstable interactions or possible switches. This is shown for a methionine residue of 4ZYZ (upper panel) and a tryptophan residue of 1WAW (lower panel). Coloring as in b after 2 ns and 10 ns. e , MD simulations captured local adaptability of the binding pocket and ligand. That is, in structure 2IG0 parts of the ligand (licorice, carbons in ivory) are quite flexible in the protein pocket (gray carbons) when comparing the first (dark hue) and the last (light hue) frames of the MD run. Color and character keys, if not indicated differently: N, blue, nitrogen; S, yellow, sulfur; O, red, oxygen; C, black, carbon; H, white, hydrogen; F, beige, fluorine; P, orange, phosphorus.

Typical limitations in structural datasets

Understanding the nature and sources of errors in structural databases is imperative for improving the quality of the underlying molecular models.

Macromolecule–ligand interaction strength, the most desired baseline parameter for DD, is unfortunately also the most inaccurate metric. The diverse experimental set-ups from experimental entropy/enthalpy determination (for example, isothermal titration calorimetry) to cellular phenotypic response are given as ligand strength. These values are not comparable and their use to train AI models is generally unreliable. To enable validation of affinity prediction we have prepared a small subset of ligands with accurately determined affinities to be used as a benchmark (Supplementary Table 1 ). We also tested our example model against it.

As MISATO is founded on experimental data, the two main sources of structural inaccuracies must be corrected. These are limited spatial resolution of the experimental structures and problems and biases associated with the software used for processing the molecular geometries. As well as the absence of hydrogen atoms in crystallographic structures, resolution affects the heteroatom geometry. Contracted or elongated bonds are common (Fig. 2 ). That is, most nitro groups we examined were heavily distorted: in the 1WUG structure 34 , NO bonds are almost 17% larger than reference experimental data 35 . Another example is seen in the 4MDN structure 36 , where an amide was so distorted that it explicitly violated VSEPR (valence shell electron pair repulsion) theory. Reinspection of the experimental electronic density hinted that the CÔC angle in the 4-chlorobenzyl phenyl ether moiety is also larger by almost 20° against anisole, a reference compound for that bond angle 35 . Simultaneous relaxation of the two groups leads to substantial improvement, in particular an amide group very close to reference structural values. Such errors in the heteroatom skeleton propagate further when assigning and counting hydrogen atoms. In the 5GTR structure 37 , a guanidino group strongly deviates from the expected planarity. The immediate consequences are incorrect atomic hybridizations and overassignment of hydrogen atoms, with a local formal charge of +3 in a radius of one bond around the central carbon. More examples are described in Supplementary Information .

Evaluation of the QM-based ligand curation

Employing the protocol defined in Supplementary Section 6 we modified a total of 3,930 structures, which corresponds roughly to 20% of the original database that needed substantial refinement (Fig. 2 ). Of these, 3,905 cases involve changes in protonation states, while changes in heteroatoms involve 97 ligands. These are predominantly the addition of model functional groups to emulate covalent binding with the protein (20) or the addition of missing hydroxyl groups to boronic acids.

Some ligands were split into several molecules as the original structures were not binary protein–ligand complexes (one ligand): 1A0T, 1G42, 1G9D, 2L65, 3D4F and 4MNV. 1E55 is supposed to be a mixture of two entities. However, the closest contact between them is insufficient to consider them separately, but also too large for a covalent interaction. Similar considerations apply to 1F4Y, though here close intramolecular contacts are at stake. In 4AW8 we observed a substantial deformation for the published ligand, PG6. We observed that the reference affinity is related to the metal ion in the system, Zn( ii ), and not to PG6. The structure was consequently excluded.

As depicted in Fig. 2 , the most common adjustment was the removal of hydrogen atoms from the initial PDBbind geometry. This amounts to almost 75% of the modifications. It has been pointed out that libraries such as PDBbind possess biased datasets in terms of binding configurations 31 .

QM-derived properties

We calculated several molecular and atomic properties for the ligands (Supplementary Table 2 ). For the former, we include electron affinities, chemical hardness, electronegativity, ionization potentials (by definition and using Koopmans’ theorem), static log P and polarizabilities. The latter were obtained in vacuum, water and wet octanol. Atomic properties include partial charges from different models, atomic polarizabilities, bond orders, atomic hybridizations, orbital- and charge-based reactivity (Fukui) indices and atomic softness. Reactivity indices and atomic softness are derived for interactions with electrophiles, nucleophiles and radicals. Finally, we also provide tight-binding electronic densities for all ligands. Partial charges were calculated at several levels, as these are somewhat method-sensitive quantities. AM1 charges are usually the starting point for charge-correcting schemes to be used in MD simulations. This is the case for AM1-BCC 38 . Taking our AM1 charges and multiplying them by 1.14 (in the case of neutral molecules) yields 1.14*CM1A-LBCC charges 39 used in OPLS-AA simulations 40 . The main advantage of the charges we provide is that these were obtained, when required, with a HOMO (highest occupied molecular orbital)–LUMO (lowest unoccupied molecular orbital) level shift to ensure convergence to sensible electronic states. Beyond MD simulations, CMx charges 41 , 42 , 43 have also been shown to provide good estimates of molecular dipole moments, just like tight-binding Mulliken charges 44 . From the latter, we infer furthermore the reasonableness of the electronic densities provided.

MD simulations

Experimental structural data are static snapshots that are assumed to represent a thermodynamic most stable state trapped in a crystal but ignore the presence of conformational dynamics. Experimental description of dynamics in biological macromolecules from nanosecond to millisecond timescales is challenging and requires a combination of different spectroscopic techniques. NMR spectroscopy and fluorescence-based methods can provide relevant information but are time consuming, and so far the dynamic information is not well captured in public databases. MD simulations can be performed, starting from experimental structures, and letting them evolve in time using a force field that describes the molecular potential energy surface. Typically, periods of nanoseconds to microseconds can be achieved for individual systems, depending on system size. MD traces allow the analysis of small-range structural fluctuations of the protein–ligand complex, but in some cases large-scale rare events can be observed (Fig. 3 ). In existing DD software these events are mostly neglected. MD simulations of 16,972 protein–ligand complexes in explicit water were performed for 10 ns. Structures were disregarded whenever non-standard ligand atoms or inconsistencies in the protein starting structures were encountered. A variety of metadata were generated from the simulations to facilitate future AI learning (Fig. 4 , Supplementary Table 2 and Supplementary Fig. 1 ). RMSD Ligand (root-mean-square deviation of the ligand after alignment of the protein) and the root-mean-square deviation of the whole complex were calculated with respect to the native structure. Also, binding affinities were estimated using molecular mechanics generalized Born surface area (MMGBSA) scoring (no entropic contributions explicitly considered) 45 . Moreover, the buried solvent accessible surface area was obtained for the complex. Calculated properties are stable over the simulations, proving them well equilibrated (Supplementary Fig. 1 ). For some systems, larger rearrangements of the binding site were captured that in extreme cases led to an opening of the whole binding pocket (Fig. 3 ). These rare events indicate possible cryptic pockets or transient binding modes. In a small fraction of cases, dissociation was detected (details given in Supplementary Fig. 2 ).

The QM data can be accessed via the PDB ID. The properties are split by atom properties and molecular properties. Examples of the calculated molecular properties are given. The electronic densities are provided in a separate file. The MD data are also subdivided by PDB ID. The properties are calculated either for all atoms, for each timestep (frame), or for the whole trajectory, as indicated by the name.

To exemplify possible applications of our dataset, baseline AI models were trained and evaluated. These are included in the repository as a template for future community development. For the QM dataset, the electron affinity and the chemical hardness of the ligand molecules were predicted (Fig. 5 ). The Pearson correlation is 0.75 for electron affinity and 0.77 for chemical hardness. The mean absolute error shows close predictions to the target values: on average 0.12 eV for electron affinity and 0.13 eV for chemical hardness. For these two exemplary QM features, high accuracy was achieved, opening a route to a fast derivation of QM properties. This is particularly important for larger molecules, where long calculation times are frequent.

a , Scatter plot of the predicted against target values of chemical hardness and electron affinity. The AI baseline models to predict QM properties have a high correlation of 0.75 and 0.77 for electron affinity and chemical hardness, respectively. b , Adaptability is a measure of the per-atom conformational plasticity of the protein. A histogram of the correlation and the correct top 100 predictions of the adaptability for all structures in the test set are given. An overall mean correlation of 0.66 can be achieved and the mean top 100 accuracy was 0.42 for the adaptability predictions (MD). c , Scatter plot for the adaptability result (as in a ) of example structure 2IG0. The predicted values are more narrowly distributed than the actual values, but the general trend is correct, as shown by a high correlation value of 0.75. d , The adaptability of the residues in the protein pocket highly deviates between the amino acids. The AI model predicts the adaptability given in blue-shaded (target) and red-shaded (AI-predicted) spheres. The radius is scaled according to the adaptability value. The model can correctly identify the rigid residues (small spheres) but also the amino acids with high flexibility. Color and character keys: N, blue, nitrogen; S, yellow, sulfur; O, red, oxygen; C, beige for ligand atoms and black for protein atoms, carbon.

For the MD traces, the induced-fit capability of the protein (adaptability) was predicted (see Methods for an exact definition). The model was able to identify elements of biomolecule structure likely to adapt to ligand binding. We achieved a mean Pearson correlation of 0.66. On average 42 of the top 100 atoms were correctly predicted (Fig. 5 ). As shown in Fig. 5d , the model can predict the atoms in the protein pocket that are mostly flexible during the MD run (large spheres), and detect the more rigid protein regions (small spheres). This allows a fast examination of the protein pocket without the necessity of a lengthy MD setup and simulation. The adaptability model gives an innovative example of how experimental structures can be enhanced from the MD-based MISATO data.

A binding affinity AI model combines MISATO MD and QM data. Experimental binding affinities are known to be difficult to compare across different experimental techniques, experimental conditions and calculated affinity types. To decrease these effects, our affinity model predicts a relative affinity of a target structure in relation to a defined base complex. These pairs have the protein and affinity type in common. We achieved high correlations for the MISATO binding affinity benchmark, with improved results using MISATO features when compared with no MISATO features (Fig. 6 ).

a , Spearman correlation of the affinity GNN model on the binding affinity benchmark including MISATO features and without features. Moreover, the results using Vina and non-curated complexes (original PDBbind) are shown. We achieved a consistently better performance including QM charges and MD adaptabilities as MISATO features across the affinity benchmark when compared with all other approaches. b , Histogram of the correlation of experimental B factors from X-ray crystallography experiments with RMSF calculations from the MD simulations in MISATO. A correlation of 0.59 over all structures was achieved. c , High correlations of calculated Koopmans ionization potentials (IP) from ULYSSES with DFT ionization potentials (upper panel) and experimental oxidation potentials (middle and lower panels) were found for different molecule families. d , The cap-binding domain of influenza virus polymerase as a model system for experimental validation of the predicted adaptability. Values given by our AI model had a high correlation of 0.63 against the experimentally determined B factors (which, despite characterizing atom thermal vibration, usually indicates flexibility). e , Results of the hetNOE experiments of the cap-binding domain of influenza virus polymerase indicating flexibility of the protein chain were in high accordance with the results of the adaptability model (indicated using shaded regions).

Experimental validation

The MISATO database and the adaptability AI model were validated on experimental data (Fig. 6 ). In X-ray crystallography a B factor is determined for each structure in the PDB. It is a measure of the thermal vibration of each atom but usually reflects localized molecular motion as well 46 . We achieved a mean correlation of 0.59 of the B factors with the root-mean-square fluctuation (RMSF) in the MISATO MD trajectories. To prove the model against more direct experimental flexibility data, we measured the cap-binding domain of influenza virus polymerase subunit PB2 47 as a model system. Heteronuclear Overhauser effect (hetNOE) NMR measurements, which elucidate flexible protein regions in solution, were performed on this structure (Supplementary Fig. 3 ). We obtained a high correlation between the calculated adaptabilities and both B factors (0.63) and the hetNOE of the protein. A comparison with our adaptability prediction shows that the most flexible regions and the residues of higher rigidity are correctly identified by the model. Quantum chemical methods are required to predict reasonable values for ionization potentials and electron affinities 48 . This applies not only to DFT (density functional theory) but also to ab initio. In Supplementary Data 1 we provide a parameter study performed with data collected from the CCCBDB database 35 , verifying the generality of trends reported in the literature 48 , 49 . The parameter study shows furthermore that semi-empirical ionization potentials are of a quality similar to, if not higher than, the best DFT results. The advantage, however, is that we systematically apply the same level of theory for all molecules, small and very large alike. We validated the ULYSSES-based calculations of Koopmans ionization potentials against experimental oxidation potentials and DFT-based ionization potentials from the literature for three molecule families. Our calculations correlated highly for photocatalysts (0.84, experimental oxidation potential), catechins (0.68, experimental oxidation potential) and thiaflavans (0.97, calculated DFT data).

Binding affinity benchmark and validation

The numerical values describing ligand potency cannot serve as a reliable baseline due to their origin in a wide range of experiments and conditions. These errors in the ground truth cannot be averaged out efficiently. Therefore, we collected high-quality affinity data for 127 ligands for five different protein structures for a MISATO binding affinity benchmark set (Supplementary Table 1 and Supplementary Data 2 ) 50 , 51 , 52 , 53 , 54 . This set, being too small for training, can be a reliable validation method for affinity-predicting models. To guarantee reliable affinity data we filtered it to originate from the same publication for each set. Moreover, each of the sets had at least 15 entries with a high dynamical range and few additional occurrences of the protein structure within MISATO.

Our binding affinity graph convolutional network model was evaluated on this benchmark set with and without MISATO features. Additionally, we evaluated the model performance on the original PDBbind set, and using the Vina scoring function. With MD-derived adaptabilities and QM charges, we obtained a mean Spearman correlation of 0.64, which was higher than without the MISATO features (0.50), using Vina (0.51) and using the non-curated database (0.50). Interestingly, an improvement for each of the five sets using the MISATO features could be achieved.

As confirmation of the given results, we evaluated the affinity model on a second benchmark set comprising the six largest clusters of protein structures (clustered on the basis of UniProt ID) of the test set (Supplementary Table 3 ). These clusters are substantially larger than the sets from the MISATO benchmark and do not necessarily originate from the same publication and the same experimental method within a set. The absolute correlations decreased for this second, more diverse benchmark (Supplementary Fig. 4 ). Still, we see the same trend as for the first benchmark with a better performance of the MISATO model including adaptability and QM features than the other approaches.

Finally, consistent improvement of affinity prediction model accuracy upon inclusion of QM and dynamic features was observed for the entire curated set as well as selected subsets with high-confidence affinity values (Fig. 6 and Supplementary Figs. 4 and 5 ). This emphasizes the importance of curation of ligand data and inclusion of at least short-term dynamics in the accuracy of affinity predictions.

The given experiments show that adding the features present in MISATO improves model accuracy over relying on implicit learning from the bare structure.

The great advances over the past years of AI technologies were only possible due to the huge datasets that are fed into these models. In structural biology, the protein folding problem was solved recently, but the DD community still lacks a breakthrough model.

Here, we present MISATO, a database that will open routes in DD for researchers from chemistry, structural biology, biophysics and bioinformatics. MISATO contains the quantum-chemically refined ligand dataset, which permitted the elimination of several structural inaccuracies and crystallographic artifacts. Our refinement protocol can be immediately applied by others for quick database augmentation. We enhance the curated dataset following two orthogonal dimensions. On the one hand, a QM approach supplies systematic electronic properties. On the other hand, a classical approach reveals the system’s dynamics and includes the binding affinity and conformational landscape. MISATO contains the largest collection of protein–ligand MD traces to date. Extensive experimental validation of the QM calculations, MD trajectories and AI baseline models highlights the dataset’s importance (Fig. 6 ).

Checkpoint files are made available for potential community extension of the dynamic traces (Supplementary Table 4 ). Structural biology datasets until now have been unable to incorporate entropy-related information about binding sites and the dynamics of the systems. By conducting MD simulations, it is possible to approximate the conformational space for entropy estimation. A Python interface, built to be intuitively used by anyone, provides preprocessing scripts and template notebooks.

The current limitations of MISATO include the fact that until now the QM calculations were only conducted on the ligand molecules. Moreover, longer timescales of the MD simulations are desirable. These limitations are related to the availability of computing resources. With future releases of MISATO these points will be addressed.

The dataset augmentation presented here paves the way for creative applications of AI models. Our example graph neural network (GNN) model offers quick access to pocket flexibility, a problem never tackled before. This is however just a starting point for a whole class of AI models sprouting from MISATO. Ultimately, we envision models being built on the best of quantum and Newtonian worlds to obtain high-quality thermodynamics, innovatively and efficiently matching the quality of experimental data. With MISATO, AI models will uncover hidden state variables describing protein–ligand complexes.

Altogether, MISATO is meant to provide sufficient training power for accurate, next-generation structure-based DD using AI methods.

Semi-empirical calculations

QM calculations were performed using the ULYSSES library 55 , our in-house semi-empirical package. The methods of choice were GFN2-xTB 56 , AM1 (ref. 57 ) and PM6 (ref. 58 ). Implicit solvation was included using ALPB 59 as parameterized for GFN2-xTB. Selected media included water and wet octanol. Bond orders and hybridizations were estimated using distance-based criteria.

QM curation of ligand space

Consistent atomic assignments were determined using a series of semi-empirical tests. Semi-empirical quantum chemical methods offer a good compromise between accuracy and computational efficiency 60 , which is suitable to refine a collection of almost 20,000 structures of various chemical natures and dimensions (from 6 to almost 370 atoms per molecule). The consistency tests we designed were performed in vacuum to ensure maximum sensitivity of the calculations to structural inconsistencies. Predicted properties, however, are also obtained using an implicit solvation model.

It is well documented that molecules with many polar groups lack convergence in wavefunction optimization 61 . The same applies when incorrect charges or protonation states are used. Implicit solvation substantially ameliorates the issue and masks problems. In fact, after determining the first guess for total molecular charges, single-point-energy calculations on unrefined ligands using implicit water required roughly 6 h of computation time. Turning off implicit solvation increased the calculation time to almost three weeks on the same machine. This was indicative of severe limitations in proton and total charge assignment. Alternative protonation algorithms were tested—for example, Open Babel 62 . Due to experimental inaccuracies in the geometries, the results were still faulty (Supplementary Figs. 6–9 ).

Our refinement protocol started with a search for structures with strong atomic overlap. Next, we looked for structures with problematic wavefunction convergence. Vanishing HOMO–LUMO gaps or unpaired electrons flagged further problems, as did violations of the octet rule based on QM population analysis. Finally, we searched for changes in ligand connectivity patterns after QM geometry optimization. This was particularly useful in determining inconsistent protonation states or incorrect electron counting, which generated biradicals. Calculated properties yielded additional testing grounds. Incorrect element assignments were detected when plotting the partial charges against D4 polarizabilities 63 (Fig. 2b ).

Severe structural deformations were also detected, inconsistent with the chemical structure (see previous section). For the current stage of the database, we decided to fix only the most extreme cases. This was done using Avogadro (Supplementary Fig. 10 ) 64 . Further structural refinement is planned.

Whenever our corrections seemed questionable, or the structure was unclear, we checked the original publication. Oxidation states were another sensible point for ligands containing transition metals. Examples of structures we refined are given in the Supplementary Information (Supplementary Figs. 10–12 ). To ease the inclusion and processing of new structures, a heuristics-based program is included in the database, which performs the basic structural processing (see Supplementary Information for more details). A detailed schematic for the protocol used for cleaning and refining the structures is also given in the Supplementary Information (Supplementary Figs. 13 and 14 ).

For all MD simulations, we used the Amber20 (ref. 65 ) software suite. The protein–ligand complexes were prepared and simulated on the basis of a standard set-up. We parameterized the ligands calculating AM1-BCC 38 charges using antechamber 66 (if the charges did not converge within 1 h we used AM1 charges calculated with ULYSSES). We used the gaff2 (ref. 66 ) force field for ligands and ff14SB 67 for the proteins. The complexes were neutralized with Na + and Cl − ions and solvated in TIP3P 68 explicit water using periodic boundary conditions in an octahedral box (minimum distance between protein and boundary 12 Å).

The complexes were minimized (1,000 steps steepest descent followed by conjugate gradient) and heated to 300 K in several steps within 16 ps. We performed production simulations for 10 ns on all protein–ligand cases in an NVT ensemble. The first 2 ns were discarded as equilibration phase, so 8 ns are stored over 100 snapshots for each protein–ligand complex. Using pytraj 31 we calculated different properties of the simulations such as the MMGBSA interaction energy, the buried solvent accessible surface area, the center-of-mass distance between ligand and receptor, and root-mean-square deviations from the native complex.

Access to the database

The database can be downloaded from Zenodo (Supplementary Table 4 ). Data are stored in a hierarchical data format. We created two H5 files, one for the protein–ligand dynamics and one for quantum chemical data, that can be accessed through our container images or after installation of the required Python packages. Installation instructions are given in the repository (Supplementary Table 4 ). Data are split for each structure using the PDB ID. The feature of interest must also be specified (Fig. 4 and Supplementary Table 2 ). Python scripts are given in the repository showing how to preprocess the MD dataset for specific cases, only C α atoms, no hydrogen atoms, only atoms from the binding pocket, and the inclusion of new features. Instructions on how to run inference on new PDB files and visualize the baseline models are given. Checkpoint files for continuing the MD simulations and the electronic densities are provided separately.

AI applications

For the baseline model for QM predictions, we followed the GNN architecture for small-molecule property prediction in ATOM3D 69 . This model is based on graph convolutions proposed by Kipf and Welling 70 and was adapted for the simultaneous prediction of electron affinity and chemical hardness as essential parameters to describe the ligand. The architecture for the baseline model was a dense layer followed by three sequential layers of NNConv and GRU followed by two dense layers. The model is available via our GitHub repository.

The performance of the ML model was evaluated using correlation and the mean absolute error.

We encode each molecule using the atom positions, the atom type and the bond between the atoms. Each atom corresponds to a node. The atom types are one-hot encoded and edges are defined by selecting the nearest neighbors with a distance of 4.5 Å for each atom. Edges are weighted inversely by the distance between the atoms. We removed outliers straying more than 20 s.d. from the mean values (PDB IDs given in Supplementary Information ). All outliers corresponded to molecules containing negatively charged groups and alkyl chains. In other words, these are highly saturated molecules from the electronic viewpoint. Because of their electronic structure, acceptance of an electron is highly unlikely, resulting in very low-to-negative electron affinities. Inaccuracies in the geometries further exacerbate the calculated electron affinities. The results on these systems indicate that some electronic properties are not quantitative; instead, they simply reflect the system’s behavior. We trained the GNN with four NVIDIA A100 graphics processing units (GPUs) and 96 CPUs (from 48 physical cores) and for 200 epochs. We used a batch size of 128 and applied a random translation on each node of 0.05 Å.

For the MD task, we modified the GNN architecture from ATOM3D 69 for the node regression task by removing the aggregation of node features into graph features. The architecture for the baseline model was five sequential GCNConv layers 70 followed by two linear layers, summing to 370,000 trainable parameters. The dataset was split into a train (80%), a test (10%) and a validation set (10%) (Supplementary Table 5 and Supplementary Fig. 15 ) by clustering the amino-acid sequences of the proteins using BlastP 71 to make sure to not have a leakage of similar structural motifs between the splits. We train the GNN with four NVIDIA A100 GPUs and 96 CPUs and for 15 epochs. We use a batch size of eight and a random translation of 0.05 Å. With our model, we calculated the adaptability of each atom during the MD simulation. To this end, we performed an alignment of the coordinates of each simulation with reference to the first frame. To calculate the adaptability γ x for each atom x we take the mean distance of each atom over all timesteps i from the initial position of the atom r ref,x :

Hydrogen atoms were omitted to reduce the size of the model. For the evaluation, the mean over the results for each structure was calculated. Adaptability gives results very similar to those of RMSF evaluations. We evaluated the performance of our training using Pearson correlation and the average accuracy of the 100 most flexible atoms of each complex.

For the binding affinity task, the data processing, training procedure and GNN architecture were modified. For data processing, all protein–ligand complexes (excluding 1,192 protein–peptide complexes) with known binding affinity were clustered at 30% sequence similarity to avoid data leakage between training (82%), validation (9%) and test (9%) sets (Supplementary Fig. 15 ). The MISATO affinity benchmark was a holdout part of the test set. Next, clusters were defined on the basis of the UniProt identifier and affinity type, so that each cluster contained only affinity values of the same protein and one of the three affinity types present in the dataset ( K i , K d , IC50).

The model predicts the ratio of binding affinities between a pair of protein–ligand complexes. For each cluster, one base molecule was defined that built a pair with each entry of the cluster. The protein–ligand complexes for which no cluster with at least two entries could be defined were discarded (2,259 entries). The atom types were one-hot encoded (omitting hydrogen atoms), and edges were defined following the adaptability model.

One training step consisted of one forward pass for each of the two complexes and mean squared error loss calculation based on the logarithmic ratio of the affinities for each pair. We trained a model including MISATO features and without MISATO features. MISATO atom features comprised calculated adaptabilities (MD) and GFN2-xTB charges in water (QM) for the ligands.

For the GNN architecture, five sequential GCNConv layers were followed by a separate pooling operation for the ligand and protein, respectively. These representations were then further processed via three linear layers with ReLU nonlinearities.

We trained the GNN with four GPUs, 90 CPUs, a batch size of 50 and for 50 epochs. We evaluated the best models on the MISATO affinity benchmark using Spearman correlation on each set.

We used PyTorch v.1.14 to train the models. To code the data loaders and the GNN, we used PyTorch Geometric 2.3.0.

Scoring of ligands with AutoDock Vina

We calculated an AutoDock Vina 9 score for the MISATO refined protein–ligand complexes of both benchmark sets. We followed a standard preprocessing procedure of generating pdbqt files (see ref. 72 to follow the exact steps). For the receptors we used the prepare_receptor tool on the protonated protein structure from ADFR Suite 73 , 74 . For the ligands we converted the structures from MOL2 format to pdbqt format using the mk_prepare_ligand.py script. We computed the Vina scores from the generated pdbqt files using the score function from the Python interface of Vina (all scripts can be found via the GitHub page of Vina).

Correlation with experimental B factors

The experimental B factors were parsed from published PDB files of crystal structures. For data cleaning, we omitted structures for which 80% of the published B factor values had the same entry. Additionally, for some structures, it was not possible to parse the B factors correctly due to inconsistencies in the underlying PDB files. The RMSF of each atom of the MD simulation was calculated after superposition to the first frame using pytraj 31 .

Binding affinity benchmark

The benchmark was created by identifying structures in MISATO that originated from the same publication with at least 15 entries. The benchmark was carefully evaluated by assessment of the publication for each of the sets. Only high-quality experimental techniques and data were considered. We further removed sets with a small dynamic range of the affinity data, high coexistence of structures of the same protein within MISATO, and sets with cofactors or metals interacting at the binding site. We obtained a benchmark consisting of five protein sets and 127 bound ligands.

Protein purification and NMR spectroscopy

The influenza PB2 domain was expressed and purified as previously published 47 . NMR data were acquired at 298 K using a 0.8 mM 13 C- 15 N-PB2 sample on an AV600 spectrometer equipped with a cryoprobe. The sample buffer contained 20 mM sodium phosphate at pH 6.5, and 100 mM NaCl. Standard NMR experiments were used for chemical shift assignments, mainly HNCA, HNCACB, CBCACONH, HNCO, CCONH and HCCH/TOCSY (total correlation spectroscopy). Spectra were processed with the nmrDraw/NMRPipe package 75 and analyzed with NMRView 76 .

Statistics and reproducibility

The splits for train, test and validation were randomized for the different ML models. The exact procedure for each model is given in Supplementary Fig. 15 . No statistical method was used to predetermine sample size. For MD, structures were disregarded whenever non-standard ligand atoms (metal ions) or inconsistencies in the protein starting structures were encountered. For the QM model (Supplementary Section 3 ), a small number (30) of structures were omitted due to the inability of the current algorithm to provide correct predictions for them. This does not introduce a bias to the observation and does not change our observations.

The investigators were not blinded to allocation during experiments or outcome assessment.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

MISATO is publicly accessible and can be downloaded from Zenodo 77 ( https://zenodo.org/records/7711953 ). We provide instructions for usage, data loaders via our GitHub repository, and a container image with all relevant packages installed for GPU usage (Supplementary Table 4 ). MISATO was built from the PDBbind database (release 2022). Source Data are provided with this paper.

Code availability

The code can be accessed from our GitHub repository and on Zenodo 78 ( https://github.com/t7morgen/misato-dataset ). The dataset is accessible via a Python interface using a simple PyTorch data loader. Special attention was given to code modularity, which makes it easy to adjust the AI architecture (Fig. 4 and Supplementary Section 7 ). We have implemented our dataset according to the ATOM3D 69 code base, a comprehensive suite of ML methods for molecular applications.

Jumper, J. et al. Highly accurate protein structure prediction with AlphaFold. Nature 596 , 583–589 (2021).

Article Google Scholar

Berman, H., Henrick, K. & Nakamura, H. Announcing the worldwide Protein Data Bank. Nat. Struct. Mol. Biol. 10 , 980 (2003).

Mohs, R. C. & Greig, N. H. Drug discovery and development: role of basic biological research. Alzheimer’s Dement. Transl. Res. Clin. Interv. 3 , 651–657 (2017).

Sliwoski, G., Kothiwale, S., Meiler, J. & Lowe, E. W. Computational methods in drug discovery. Pharm. Rev. 66 , 334–395 (2014).

Thiel, W. Semiempirical quantum-chemical methods. WIREs Comput. Mol. Sci. 4 , 145–157 (2014).

Hollingsworth, S. A. & Dror, R. O. Molecular dynamics simulation for all. Neuron 99 , 1129–1143 (2018).

Siebenmorgen, T. & Zacharias, M. Computational prediction of protein–protein binding affinities. WIREs Comput. Mol. Sci. 10 , e1448 (2020).

Trott, O. & Olson, A. J. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J. Comput. Chem. 31 , 455–461 (2010).

Kmiecik, S. et al. Coarse-grained protein models and their applications. Chem. Rev. 116 , 7898–7936 (2016).

Spicher, S. & Grimme, S. Robust atomistic modeling of materials, organometallic, and biochemical systems. Angew. Chem. Int. Ed. 59 , 15665–15673 (2020).

Vandenbrande, S., Waroquier, M., Speybroeck, V. V. & Verstraelen, T. The monomer electron density force field (MEDFF): a physically inspired model for noncovalent interactions. J. Chem. Theory Comput. 13 , 161–179 (2017).

Wang, J. & Dokholyan, N. V. Yuel: improving the generalizability of structure-free compound–protein interaction prediction. J. Chem. Inf. Model. 62 , 463–471 (2022).

Ponder, J. W. et al. Current status of the AMOEBA polarizable force field. J. Phys. Chem. B 114 , 2549–2564 (2010).

Chen, B. et al. Automated discovery of fundamental variables hidden in experimental data. Nat. Comput Sci. 2 , 433–442 (2022).

Durrant, J. D. & McCammon, J. A. NNScore: a neural-network-based scoring function for the characterization of protein−ligand complexes. J. Chem. Inf. Model. 50 , 1865–1871 (2010).

Wang, X., Terashi, G., Christoffer, C. W., Zhu, M. & Kihara, D. Protein docking model evaluation by 3D deep convolutional neural networks. Bioinformatics 36 , 2113–2118 (2020).

Wang, N.-N. et al. ADME properties evaluation in drug discovery: prediction of Caco-2 cell permeability using a combination of NSGA-II and boosting. J. Chem. Inf. Model. 56 , 763–773 (2016).

Ishida, S., Terayama, K., Kojima, R., Takasu, K. & Okuno, Y. AI-driven synthetic route design incorporated with retrosynthesis knowledge. J. Chem. Inf. Model. 62 , 1357–1367 (2022).

Karpov, P., Godin, G. & Tetko, I. V. A transformer model for retrosynthesis. In Artificial Neural Networks and Machine Learning—ICANN 2019: Workshop and Special Sessions (eds Tetko, I. V. et al.) 817–830 (Springer, 2019).

Öztürk, H., Özgür, A. & Ozkirimli, E. DeepDTA: deep drug–target binding affinity prediction. Bioinformatics 34 , i821–i829 (2018).

Karimi, M., Wu, D., Wang, Z. & Shen, Y. DeepAffinity: interpretable deep learning of compound–protein affinity through unified recurrent and convolutional neural networks. Bioinformatics 35 , 3329–3338 (2019).

Hassan-Harrirou, H., Zhang, C. & Lemmin, T. RosENet: improving binding affinity prediction by leveraging molecular mechanics energies with an ensemble of 3D convolutional neural networks. J. Chem. Inf. Model. 60 , 2791–2802 (2020).

Feinberg, E. N. et al. PotentialNet for molecular property prediction. ACS Cent. Sci. 4 , 1520–1530 (2018).

Li, Y., Rezaei, M. A., Li, C. & Li, X. DeepAtom: a framework for protein–ligand binding affinity prediction. In 2019 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) 303–310 (IEEE, 2019).

Wang, R., Fang, X., Lu, Y., Yang, C.-Y. & Wang, S. The PDBbind database: methodologies and updates. J. Med. Chem. 48 , 4111–4119 (2005).

Liu, T., Lin, Y., Wen, X., Jorissen, R. N. & Gilson, M. K. BindingDB: a web-accessible database of experimentally determined protein–ligand binding affinities. Nucleic Acids Res. 35 , D198–D201 (2007).

Hu, L., Benson, M. L., Smith, R. D., Lerner, M. G. & Carlson, H. A. Binding MOAD (Mother Of All Databases). Proteins Struct. Funct. Bioinform. 60 , 333–340 (2005).

Friedrich, N.-O., Simsir, M. & Kirchmair, J. How diverse are the protein-bound conformations of small-molecule drugs and cofactors? Front. Chem. 6 , 68 (2018).

Korlepara, D. B. et al. PLAS-5k: dataset of protein–ligand affinities from molecular dynamics for machine learning applications. Sci. Data 9 , 548 (2022).

Korlepara, D. B. et al. PLAS-20k: extended dataset of protein–ligand affinities from MD simulations for machine learning applications. Sci. Data 11 , 180 (2024).

Yang, J., Shen, C. & Huang, N. Predicting or pretending: artificial intelligence for protein–ligand interactions lack of sufficiently large and unbiased datasets. Front. Pharmacol. 11 , 69 (2020).

Volkov, M. et al. On the frustration to predict binding affinities from protein–ligand structures with deep neural networks. J. Med. Chem. 65 , 7946–7958 (2022).

Vajda, S., Beglov, D., Wakefield, A. E., Egbert, M. & Whitty, A. Cryptic binding sites on proteins: definition, detection, and druggability. Curr. Opin. Chem. Biol. 44 , 1–8 (2018).

Zeng, L. et al. Selective small molecules blocking HIV-1 Tat and coactivator PCAF association. J. Am. Chem. Soc. 127 , 2376–2377 (2005).

Johnson, R. D. III (ed). Computational Chemistry Comparison and Benchmark Database Standard Reference Database Number 101 Release 22 (NIST, accessed 12 Jul 2022); http://cccbdb.nist.gov/

Bista, M. et al. Transient protein states in designing inhibitors of the MDM2–p53 interaction. Structure 21 , 2143–2151 (2013).

Xie, M. et al. Structural basis of inhibition of ERα–coactivator interaction by high-affinity N -terminus isoaspartic acid tethered helical peptides. J. Med. Chem. 60 , 8731–8740 (2017).

Jakalian, A., Jack, D. B. & Bayly, C. I. Fast, efficient generation of high-quality atomic charges. AM1-BCC model: II. Parameterization and validation. J. Comput. Chem. 23 , 1623–1641 (2002).

Dodda, L. S., Vilseck, J. Z., Tirado-Rives, J. & Jorgensen, W. L. 1.14*CM1A-LBCC: localized bond-charge corrected CM1A charges for condensed-phase simulations. J. Phys. Chem. B 121 , 3864–3870 (2017).

Jorgensen, W. L., Maxwell, D. S. & Tirado-Rives, J. Development and testing of the OPLS all-atom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc. 118 , 11225–11236 (1996).

Storer, J. W., Giesen, D. J., Cramer, C. J. & Truhlar, D. G. Class IV charge models: a new semiempirical approach in quantum chemistry. J. Comput. Aided Mol. Des. 9 , 87–110 (1995).

Li, J., Zhu, T., Cramer, C. J. & Truhlar, D. G. New class IV charge model for extracting accurate partial charges from wave functions. J. Phys. Chem. A 102 , 1820–1831 (1998).

Thompson, J. D., Cramer, C. J. & Truhlar, D. G. Parameterization of charge model 3 for AM1, PM3, BLYP, and B3LYP. J. Comput. Chem. 24 , 1291–1304 (2003).

Grimme, S. & Bannwarth, C. Ultra-fast computation of electronic spectra for large systems by tight-binding based simplified Tamm–Dancoff approximation (sTDA-xTB). J. Chem. Phys. 145 , 054103 (2016).

Wang, E. et al. End-point binding free energy calculation with MM/PBSA and MM/GBSA: strategies and applications in drug design. Chem. Rev. 119 , 9478–9508 (2019).

Sun, Z., Liu, Q., Qu, G., Feng, Y. & Reetz, M. T. Utility of B factors in protein science: interpreting rigidity, flexibility, and internal motion and engineering thermostability. Chem. Rev. 119 , 1626–1665 (2019).

Guilligay, D. et al. The structural basis for cap binding by influenza virus polymerase subunit PB2. Nat. Struct. Mol. Biol. 15 , 500–506 (2008).

Rayne, S. & Forest, K. Benchmarking semiempirical, Hartree–Fock, DFT, and MP2 methods against the ionization energies and electron affinities of short- through long-chain [ n ]acenes and [ n ]phenacenes. Can. J. Chem. 94 , 251–258 (2016).

Zhan, C.-G., Nichols, J. A. & Dixon, D. A. Ionization potential, electron affinity, electronegativity, hardness, and electron excitation energy: molecular properties from density functional theory orbital energies. J. Phys. Chem. A 107 , 4184–4195 (2003).

Lange, G. et al. Requirements for specific binding of low affinity inhibitor fragments to the SH2 domain of pp60Src are identical to those for high affinity binding of full length inhibitors. J. Med. Chem. 46 , 5184–5195 (2003).

Öster, L., Tapani, S., Xue, Y. & Käck, H. Successful generation of structural information for fragment-based drug discovery. Drug Discov. Today 20 , 1104–1111 (2015).

Heinzlmeir, S. et al. Chemoproteomics-aided medicinal chemistry for the discovery of EPHA2 inhibitors. ChemMedChem 12 , 999–1011 (2017).

Gaieb, Z. et al. D3R Grand Challenge 2: blind prediction of protein–ligand poses, affinity rankings, and relative binding free energies. J. Comput. Aided Mol. Des. 32 , 1–20 (2018).

Whitehouse, A. J. et al. Development of inhibitors against Mycobacterium abscessus tRNA (m1G37) methyltransferase (TrmD) using fragment-based approaches. J. Med. Chem. 62 , 7210–7232 (2019).

Menezes, F. & Popowicz, G. M. ULYSSES: an efficient and easy to use semiempirical library for C. J. Chem. Inf. Model. 62 , 3685–3694 (2022).

Bannwarth, C., Ehlert, S. & Grimme, S. GFN2-xTB—an accurate and broadly parametrized self-consistent tight-binding quantum chemical method with multipole electrostatics and density-dependent dispersion contributions. J. Chem. Theory Comput. 15 , 1652–1671 (2019).

Dewar, M. J. S., Zoebisch, E. G., Healy, E. F. & Stewart, J. J. P. Development and use of quantum mechanical molecular models. 76. AM1: a new general purpose quantum mechanical molecular model. J. Am. Chem. Soc. 107 , 3902–3909 (1985).

Stewart, J. J. P. Application of the PM6 method to modeling proteins. J. Mol. Model. 15 , 765–805 (2009).

Sigalov, G., Fenley, A. & Onufriev, A. Analytical electrostatics for biomolecules: beyond the generalized Born approximation. J. Chem. Phys. 124 , 124902 (2006).

Christensen, A. S., Kubař, T., Cui, Q. & Elstner, M. Semiempirical quantum mechanical methods for noncovalent interactions for chemical and biochemical applications. Chem. Rev. 116 , 5301–5337 (2016).

Dixon, S. L. & Merz, K. M. Fast, accurate semiempirical molecular orbital calculations for macromolecules. J. Chem. Phys. 107 , 879–893 (1997).

O’Boyle, N. M. et al. Open Babel: an open chemical toolbox. J. Cheminform. 3 , 33 (2011).

Caldeweyher, E. et al. A generally applicable atomic-charge dependent London dispersion correction. J. Chem. Phys. 150 , 154122 (2019).

Hanwell, M. D. et al. Avogadro: an advanced semantic chemical editor, visualization, and analysis platform. J. Cheminform. 4 , 17 (2012).

Case, D. A. et al. Amber 2021 (Univ. of California, San Francisco, 2021).

Wang, J., Wolf, R. M., Caldwell, J. W., Kollman, P. A. & Case, D. A. Development and testing of a general Amber force field. J. Comput. Chem. 25 , 1157–1174 (2004).

Maier, J. A. et al. ff14SB: improving the accuracy of protein side chain and backbone parameters from ff99SB. J. Chem. Theory Comput. 11 , 3696–3713 (2015).

Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W. & Klein, M. L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79 , 926–935 (1983).

Townshend, R. J. L. et al. ATOM3D: tasks on molecules in three dimensions. Preprint at https://doi.org/10.48550/arXiv.2012.04035 (2022).

Kipf, T. N. & Welling, M. Semi-supervised classification with graph convolutional networks. Preprint at https://doi.org/10.48550/arXiv.1609.02907 (2017).

Huang, Y., Niu, B., Gao, Y., Fu, L. & Li, W. CD-HIT Suite: a web server for clustering and comparing biological sequences. Bioinformatics 26 , 680–682 (2010).

Forli, S. et al. Computational protein–ligand docking and virtual drug screening with the AutoDock suite. Nat. Protoc. 11 , 905–919 (2016).

Zhao, Y., Stoffler, D. & Sanner, M. Hierarchical and multi-resolution representation of protein flexibility. Bioinformatics 22 , 2768–2774 (2006).

Ravindranath, P. A., Forli, S., Goodsell, D. S., Olson, A. J. & Sanner, M. F. AutoDockFR: advances in protein–ligand docking with explicitly specified binding site flexibility. PLoS Comput. Biol. 11 , e1004586 (2015).

Delaglio, F. et al. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J. Biomol. NMR 6 , 277–293 (1995).

Johnson, B. A. & Blevins, R. A. NMR View: a computer program for the visualization and analysis of NMR data. J. Biomol. NMR 4 , 603–614 (1994).

Siebenmorgen, T. et al. MISATO—machine learning dataset for structure-based drug discovery. Zenodo https://doi.org/10.5281/zenodo.7711953 (2023).

t7morgen/misato-dataset: release for publication. Zenodo https://doi.org/10.5281/zenodo.10926008 (2024).

Download references

Acknowledgements

This work received funding from BMWi ZIM KK 5197901TS0 (T.S., F.M., G.M.P.) and BMBF, SUPREME, 031L0268 (T.S., F.M., G.M.P.). This work was supported by the Helmholtz Association’s Initiative and Networking Fund on the HAICORE@FZJ partition.

The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.

Open access funding provided by Helmholtz Zentrum München - Deutsches Forschungszentrum für Gesundheit und Umwelt (GmbH).

Author information

These authors contributed equally: Till Siebenmorgen, Filipe Menezes.

Authors and Affiliations

Molecular Targets and Therapeutics Center, Institute of Structural Biology, Helmholtz Munich, Neuherberg, Germany

Till Siebenmorgen, Filipe Menezes, André Santos Dias Mourão, Michael Sattler & Grzegorz M. Popowicz

TUM School of Natural Sciences, Department of Bioscience, Bayerisches NMR Zentrum, Technical University of Munich, Garching, Germany

Jülich Supercomputing Centre, Forschungszentrum Jülich, Jülich, Germany

Sabrina Benassou & Stefan Kesselheim

Helmholtz AI, Helmholtz Munich, Neuherberg, Germany

Erinc Merdivan, Marie Piraud & Fabian J. Theis

Computer Laboratory, Cambridge University, Cambridge, UK

Kieran Didi & Pietro Liò

Faculty of Chemistry, Jagiellonian University, Krakow, Poland

Radosław Kitel

Computational Health Center, Institute of Computational Biology, Helmholtz Munich, Neuherberg, Germany

Fabian J. Theis

TUM School of Computation, Information and Technology, Technical University of Munich, Garching, Germany

You can also search for this author in PubMed Google Scholar

Contributions

T.S. and F.M. created and refined the dataset, designed the ML experiments, performed the ML experiments, analyzed the data and wrote the paper. S.B., E.M. and K.D. performed the ML experiments, analyzed the data and contributed to the paper writing. A.S.D.M. performed the NMR experiments and analysis and wrote the NMR section. R.K. selected and validated the affinity benchmark. P.L., S.K., M.P., F.J.T. and M.S. contributed to study design, paper writing and funding of the project. G.M.P. conceived and designed the ML experiments, analyzed the data and wrote the paper.

Corresponding author

Correspondence to Grzegorz M. Popowicz .

Ethics declarations

Competing interests.

The authors declare no competing interests.

Peer review

Peer review information.

Nature Computational Science thanks Martin Zacharias and the other, anonymous, reviewer(s) for their contributions to the peer review of this work. Primary Handling Editor: Kaitlin McCardle, in collaboration with the Nature Computational Science team. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary information.

Supplementary Tables 1–5, Figs. 1–15 and Sections 1–7.

Reporting Summary

Peer review file, supplementary data 1.

RMS of Koopmans ionization potentials compared against experimental data, collected from CCCBDB. RMS in electronvolts.

Supplementary Data 2

Table of MISATO binding affinity benchmark. We give the PDB ID, affinity (nM and kcal mol −1 ), ligand PDB ID, SMILES, experimental method and reference for each of the five sets.

Source Data Fig. 2

Statistical source data for Fig. 2 showing electronic densities, structures, and data on halogens and for spider plot.

Source Data Fig. 3

Statistical source data for Fig. 3 showing MD trajectories for the described structures.

Source Data Fig. 5

Statistical source data for Fig. 5.

Source Data Fig. 6

Statistical source data for Fig. 6.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Siebenmorgen, T., Menezes, F., Benassou, S. et al. MISATO: machine learning dataset of protein–ligand complexes for structure-based drug discovery. Nat Comput Sci (2024). https://doi.org/10.1038/s43588-024-00627-2

Download citation

Received : 30 May 2023

Accepted : 11 April 2024

Published : 10 May 2024

DOI : https://doi.org/10.1038/s43588-024-00627-2

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

This article is cited by

A multidimensional dataset for structure-based machine learning.

Matthew Holcomb
Stefano Forli

Nature Computational Science (2024)

Quick links

Explore articles by subject
Guide to authors
Editorial policies

Sign up for the Nature Briefing: Translational Research newsletter — top stories in biotechnology, drug discovery and pharma.

Type 2 Diabetes
Heart Disease
Digestive Health
Multiple Sclerosis
Diet & Nutrition
Supplements
Health Insurance
Public Health
Patient Rights
Caregivers & Loved Ones
End of Life Concerns
Health News
Thyroid Test Analyzer
Doctor Discussion Guides
Hemoglobin A1c Test Analyzer
Lipid Test Analyzer
Complete Blood Count (CBC) Analyzer
What to Buy
Editorial Process
Meet Our Medical Expert Board

A Guide to Your Ozempic Dosage

Recommended ozempic dosages, dosage adjustments, what happens if i take too much, what should i do if i miss a dose.

Ozempic (semaglutide) is a brand-name injectable medication that belongs to a group of drugs called glucagon-like peptide 1 (GLP-1) receptor agonists .

Ozempic is approved by the Food and Drug Administration (FDA) to:

Help manage blood sugar levels in adults with type 2 diabetes , along with exercise and a healthy diet
Reduce the risk of heart and blood vessel complications, such as heart attack and stroke , in adults with type 2 diabetes and heart disease

It comes in prefilled pens that allow you to inject the medication under your skin.

The recommended starting dose of Ozempic is 0.25 milligrams (mg) once weekly. This low dose will help your body get used to the medication. After four weeks, the dose is typically increased to 0.5 mg once weekly, the maintenance dosage.

This article will review how to take Ozempic safely, including commonly recommended dosages, adjusting dosages for special circumstances, and what to do if you miss a dose.

Getty Images / Bloomberg / Contributor

Ozempic comes as a liquid that you inject under your skin . Healthcare providers typically prescribe a low starting dose. Then, they gradually increase the dosage based on how your body responds to Ozempic.

The recommended starting dose of Ozempic is 0.25 mg injected under the skin once every four weeks. This dose is not enough to effectively manage your blood sugar levels. It's used to help your body get used to digestion-related side effects, such as nausea and stomach pain while starting Ozempic. After this time, these side effects are likely to subside.

After four weeks, your healthcare provider will increase your dosage to 0.5 mg once per week. This is the recommended maintenance dose.

How to Use Ozempic

Ozempic comes in prefilled pens to inject the medication under your skin.

Inject your Ozempic dose once per week on the same day each week, with or without food. You can use Ozempic at any time during the day.

Ozempic can be injected into the abdomen, thigh, or back of the upper arm. You can inject the medication into the same or different parts of your body. If you inject each dose of Ozempic into the same part of your body, be sure to rotate the spot where you inject.

If you also use insulin with Ozempic, you may inject both medications into the same body part. However, neither injection should be given next to the other.

You should not mix insulin and Ozempic in the same syringe. Both injections should be given separately.

Your Ozempic pen may deliver more than one strength of the medication. In this case, you may need to select your dose before injecting it.

Before your first dose, your healthcare provider will likely show you how to use your Ozempic pen. They'll go over how to select the correct dose in your pen.

If you have more questions about using your Ozempic pen, talk with your pharmacist or other healthcare provider.

Dosage Forms and Strengths

Ozempic is available as three different prefilled pens. Each strength is color-coded as follows:

The red pen contains 0.25 mg and 0.5 mg doses
The blue pen contains 1 mg doses
The yellow pen contains 2 mg doses

Each red pen can deliver four 0.25 mg starting doses and two 0.5 mg maintenance doses. Your healthcare provider will likely prescribe this strength as your first pen.

The red pen can also deliver four 0.5 mg doses, so you may use this pen for your maintenance dose.

Each blue pen delivers four 1 mg doses, and the yellow pen contains four 2 mg doses. Your healthcare provider may prescribe either of these pens if your dosage needs to be increased.

The table below summarizes the dosages and strengths of each Ozempic pen.

Before you begin Ozempic treatment, your healthcare provider will likely discuss healthy blood sugar level goals with you.

They'll check your blood sugar levels four weeks after beginning your 0.5 mg maintenance dose. They may also recommend regularly monitoring your blood sugar level at home. You'll do this using a glucometer or a continuous glucose monitoring device.

If Ozempic is not effectively lowering your blood sugar, your healthcare provider may increase your dosage to 1 mg once per week.

If, after four more weeks, your blood sugar levels are still too high, they may further increase your dosage to 2 mg once per week,

The maximum dose of Ozempic is 2 mg per week.

Dosage Adjustments for Other Medications for Diabetes

Your healthcare provider may prescribe Ozempic with other medications to help manage your blood sugar levels, such as Glucotrol XL (glipizide) or insulin .

Ozempic helps to lower your blood sugar levels. Taking this medication with other drugs that also lower your blood sugar may increase your risk for hypoglycemia (very low blood sugar level).

If you take Ozempic with insulin or other drugs for diabetes, your healthcare provider may reduce the dosage of your insulin or other diabetes medications to lower your risk of hypoglycemia.

It's important to use Ozempic exactly as your healthcare provider recommends. Ozempic is a long-acting medication that stays in your body for over a week.

Using too much Ozempic can lead to serious side effects, such as hypoglycemia.

Symptoms of hypoglycemia can include:

Nervousness

Hypoglycemia symptoms may be mild at first. However, if it's left untreated, hypoglycemia may result in more serious complications, including coma and even death.

If you think you've used too much Ozempic, contact your healthcare provider right away. If it's an emergency, call 911.

If you have symptoms of hypoglycemia, eat or drink 15 grams of carbohydrates, which can quickly raise your blood sugar. Examples include:

Glucose tablets
Fruit juice
Non-diet soda
Sugary candy

If your hypoglycemia symptoms are severe enough to require help from someone else to recover, have them seek emergency medical attention immediately.

Your healthcare provider may prescribe a glucagon injection pen for you to use in case of a hypoglycemia emergency. Glucagon is a hormone that helps quickly raise your blood sugar level.

Because Ozempic stays in your system for a long time, your provider may monitor you closely for signs of hypoglycemia for a few weeks after using too much Ozempic.

It's important to use Ozempic regularly as your healthcare provider recommends.

It can take at least four weeks for Ozempic to reach steady levels in your body. Missing a dose of Ozempic may prevent the medication from working as well as it should. This may lead to an increase in your blood sugar level .

If you miss a dose of Ozempic, inject the missed dose as soon as you remember within five days of missing your dose. If it's been longer than five days since you missed your dose, skip the missed dose. Then, inject the next dose on the regularly scheduled day.

Afterward, you can continue with your regular once-per-week schedule. Do not inject two doses together to make up for a missed dose.

Ozempic is a prescription injection used along with exercise and a healthy diet to manage blood sugar levels in adults with type 2 diabetes. It is also used to reduce the risk of serious heart conditions in adults with type 2 diabetes and heart disease.

Ozempic comes in prefilled pens to inject the medication under your skin. Each pen is color-coded based on the strength of Ozempic it contains.

Dosing starts at 0.25 mg per week. This low dose is intended to help get your body used to the medication. After four weeks, the dose is typically increased to a maintenance dosage of 0.5 mg once per week.

Your healthcare provider will monitor your blood sugar levels during your Ozempic treatment. If Ozempic does not effectively lower your blood sugar, they may increase your dosage gradually every four weeks to a maximum of 2 mg once per week.

If you have questions about using your Ozempic pen, talk with your pharmacist or other healthcare provider.

National Institutes of Health. DailyMed. Label: Ozempic- semaglutide injection, solution .

Husain M, Bain SC, Holst AG, et al. Effects of semaglutide on risk of cardiovascular events across a continuum of cardiovascular risk: combined post hoc analysis of the SUSTAIN and PIONEER trials . Cardiovasc Diabetol . 2020;19(1):156. doi:10.1186/s12933-020-01106-4

Strain WD, Frenkel O, James MA, et al. Effects of semaglutide on stroke subtypes in type 2 diabetes: post hoc analysis of the randomized SUSTAIN 6 and PIONEER 6 . Stroke . 2022;53:2749–2757. doi.org/10.1161/STROKEAHA.121.037775

Food and Drug Administration. Ozempic (semaglutide) prescribing information .

Centers for Disease Control and prevention. Manage blood sugar .

PDR. Semaglutide .

Hall S, Isaacs D, Clements JN. Pharmacokinetics and clinical implications of semaglutide: a new glucagon-like peptide (GLP)-1 receptor agonist . Clin Pharmacokinet 2018;57:1529-1538. doi:10.1007/s40262-018-0668-z

MedlinePlus. Hypoglycemia .

Center for Disease Control and Prevention. How to treat hypoglycemia .

By Rosanna Sutherby, PharmD Rosanna Sutherby, PharmD, is a freelance medical writer and community pharmacist with over 20 years of experience in medication review, counseling, and immunization.

The Assignment with Audie Cornish

Every monday on the assignment, host audie cornish explores the animating forces of american politics. it’s not about the horserace, it’s about the larger cultural ideas driving the american electorate. audie draws on the deep well of cnn reporters, editors, and contributors to examine topics like the nuances of building electoral coalitions, and the role the media plays in modern elections. every thursday, audie pulls listeners out of their digital echo chambers to hear from the people whose lives intersect with the news cycle, as well as deep conversations with people driving the headlines. from astrology’s modern renaissance to the free speech wars on campus, no topic is off the table..

Apple Podcasts

Back to episodes list

What can we learn from the first great rap battle of the streaming age? Like the ones that came before, the recent rap beef between Kendrick Lamar and Drake is defined by lyrical dexterity and sheer intensity. But UNLIKE past battles this one was remarkable for its speed. Both rappers released songs within hours and even minutes of each other. Audie talks with CNN Entertainment Reporter Lisa Respers France about this moment, and how social media pushed this beef to a fever pitch.

Read Lisa’s reporting on the feud :

Kendrick Lamar and Drake gave us an epic hip-hop beef weekend. Here’s what to know

Questlove was not happy with Drake and Kendrick Lamar’s beef: ‘Nobody won the war’

© 2024 Cable News Network. A Warner Bros. Discovery Company. All Rights Reserved. CNN Audio's transcripts are made available as soon as possible. They are not fully edited for grammar or spelling and may be revised in the future. The audio record represents the final version of CNN Audio.

HHS Drops ‘Stacking’ Provision in Medicaid Drug Rebate Proposal

By Ganny Belloni

The Biden administration is stepping back from a proposal that would have changed how drugmakers calculate the “best price” they must offer Medicaid after receiving pushback from the pharmaceutical industry.

In an update released Wednesday, the Centers for Medicare & Medicaid Services said it won’t finalize a provision in the proposed rule ( RIN 0938-AU28 ) that would have required companies to “stack” discounts and rebates throughout a transaction when reporting best prices for the Medicaid drug rebate program.

Instead, the agency said it will collect more information from manufacturers on their price calculation methodologies before pursuing changes through future ...

Learn more about Bloomberg Law or Log In to keep reading:

Learn about bloomberg law.

AI-powered legal analytics, workflow tools and premium legal & business news.

Already a subscriber?

IMAGES

The nurse's quick guide to I.V. drug calculations
Drug Calculation Example 1
ASSIGNMENT 8 Chapter 4 DRUG CALCULATION PREPARING AND GIVING DRUGS DEC
Medication Dosage Calculation Nursing Study Guide
A Nurse's Ultimate Guide to Accurate Drug Dosage Calculations
ASSIGNMENT 8 Chapter 4 DRUG CALCULATION PREPARING AND GIVING DRUGS DEC

VIDEO

Drug Dose Calculation/simple and easy/ Universal formula #norcet #norcet2023 #aiims #nurseslife
DRUG CALCULATION BY MUKESH SIR JODHPUR AIIMS
IV Drug Calculation
ASSIGNMENT ON ROUTE OF DRUG ADMINISTRATION
Intravenous Calculation
Parenteral Drug Calculation Review

COMMENTS

Drug Calculations Practice NCLEX Questions (100+ Items)
Drug Dosage Calculation Practice Quiz. In this section are the practice problems and questions for nursing dosage calculations. This nursing test bank set includes 100+ questions. Included topics are dosage calculation, metric conversions, unit conversions, parenteral medications, and fluid input and output. As you can tell, this NCLEX practice ...
Dosage Calculations Assignment Quiz Flashcards
First, calculate the amount of milliliters of medication needed to provide 800 mg of nafcillin. Prescribed dose (800 mg) divided by dose on hand (250 mg/ml) equals 3.2 mlNext, determine the recommended concentration using equivalent fractions. 800 mg/X = 40 mg/1 ml40X = 800 = 20 ml (the total volume at the recommended concentration)20 ml (total ...
Dosage Calculations Assignment Quiz Flashcards
First calculate the total mg dose/day: 40 mg x 28.4 kg = 1136 mg/day. Next divide the total mg into the 4 dosages for the day: 1136 mg/4 doses = 284 mg/dose. Using the Desired/Have formula: 284 mg/50mg x 1 ml = 5.68 ml/ dose, rounded to 5.7 ml. A client weighing 86 kg is receiving dopamine at 8 mcg/kg/minute.
Pharmacology Ch. 4: Drug Calculation: Preparing and Giving Drugs
True. How should drug doses be calculated for infants? weight based, usually in kilograms. What ages use Clark's Rule to determine doses? ages 2-17. What is the equation for Clark's Rule? Pediatric child dose = (weight of child/150lbs) * adult dose. What method is the most accurate for determining pediatric dosages?
DosageHelp.com
IV Terms and Abbreviations. Amount in IV Fluid. Volume/Time - IV mL Rate. Volume/Time - IV Drop Rate. Fluid Maintenance Requirement. Dosage By Weight. Mass/Time - IV mL Rate. Practice Questions. Learn dosage calculations with this free tutorial complete with explanations, examples, and practice questions.
Dosage Calculation For Nursing Students: 15 Resources To Help You
Udemy Dosage Calculations Mastery. This $94 course is built for nurses and pharmacy students to pass their respective licensure examinations. The video course and study guide include adult and pediatric doses, understanding percentages, and commonly used medical abbreviations.
Calculating Dosages Online
Whether on a computer or tablet, Calculating Dosages Online makes learning and applying dosage calculations—even reviewing basic math principles—fun and easy! More than 2,550 questions with rationales cover exactly what you need to know for class, clinical, the NCLEX®, and professional practice.
PDF How to Solve Drug Dosage Problems
Step one: Think of a conversion factor that relates mg and gr. 60 mg = gr 1 (This can be used as either 60 mg/gr 1 or gr 1/60 mg) Step two: set up the multiplication equation. Note: when using the conversion factor, always place the needed unit on top. Step three: Solve the equation. First cancel mg units,
Dosage Calculations
2023-2024 DOSAGE CALCULATIONS ¡ Practice essential skills ¡ Build competence and confidence ¡ Prevent medication errors ¡ Ensure patient sa 1 2 20. Dosage Calculations 2023-2024 ... Turn off any assignments and topics you won't be teaching, and easily reorder assignments so they appear in the exact way you plan to teach them.
Dosage Calculation
Student-friendly interactive learning. Easy-to-master Learning Assignments deliver must-know content. Interactive readings, videos, and graphics enhance understanding of complex concepts. New! Interactive format with clickable, color-coded tabs display the math for each of the four calculation methods. Looking for more Dosage Calc resources?
Dose Calculation Ratio and Proportion Method
In colon format, you would use H:V::D:X and multiply means DV and Extremes HX. Hx = DV, x = DV/H, 2:1::4:x, 2x = (4) (1), x = 4/2, x = 2 mL. Desired Over Have or Formula Method. Desired over Have or Formula Method uses a formula or equation to solve for an unknown quantity (x), much like ratio proportion. Drug calculations require the use of ...
DosageMath.com
Welcome! DosageMath.com offers free resources to faculty and students to support the learning of medication dosage calculations.. Rather than focusing on memorizing formulas, the resources on DosageMath.com place an emphasis on understanding the "why," developing sound quantitative reasoning, and building confidence in completing dosage calculation tasks.
Dosage Calculation Practice Exam 1 Flashcards
Study with Quizlet and memorize flashcards containing terms like A patient has an order for Chloromycetin, 500 mg every 6 hours. The drug comes in 250 mg capsules. What would the nurse administer?, A nurse is to administer 150 mg of a drug intramuscularly. The label on the multidose vials reads 100 mg/mL. How much would the nurse give?, A nurse is to administer 30 mg of furosemide (Lasix) to a ...
Drug Calculations
Drug Calculations . ... It features an Elsevier ebook and interactive digital lessons with built-in assessments and adaptive quizzing assignments that allow students to learn and apply their knowledge in a more engaging way. Plus, teaching resources and insightful assignment dashboards help you track student interactions and identify ...
Medications and Calculations
Four are general methods: (1) basic formula, (2) ratio and proportion, (3) fractional equation, and (4) dimensional analysis. The nurse should select one of these general methods for the calculation of drug dosages. The other two methods are used to individualize drug dosing by body weight and body surface area.
Dosage Calculation Assignment 3
Dosage Calculation Assignment 1; Related documents. Boot Camp online; Health Assessment Mod. 5 EAQ; Health Assessment Mod. 1.2 EAQ; Clinical Pathophysiology 1 - Fall 2018; Clinical Pathophysiology 1 - Fall 2018; Narrative Documentation - Catheter - Fall 2018; Preview text.
Pharm assginment 8 Drug Calculation Preparing and Giving ...
Drug Calculation Preparing and Giving Drugs drug calculation: preparing and giving drugs assignment give the definitions to the following terminology: ampules: ... Assignment 6 & 7 Chapter 13 Drugs for Gastrointestional Problems Jan 2019; Assignment 9 CH 10 Tuberculosis RESP MEDS; Glyburide ati medication template;
Dosage Calc 360
Anatomy and Physiology ; Care Planning and Nursing Diagnoses ; Community and Public Health ; Culture ; Davis Notes for Nurses ; Electrocardiography (ECG)
Drug Dosage Calculations
Drug Dosage Calculations. Mar 22, 2022 • Download as PPTX, PDF •. 41 likes • 47,005 views. Dr. Binu Babu Nursing Lectures Incredibly Easy. Drug Dosage Calculations. Health & Medicine. Slideshow view. Download now. Drug Dosage Calculations - Download as a PDF or view online for free.
Chapter 4 Drug Calculation: Preparing and Giving Drugs
Hold the drug until clarifying the order with the prescriber. What are the steps for giving drugs to a patient? Follow the 9 Rights, minimize distractions, wash hands, assemble equipment, follow the written order, identify the patient, and do not unwrap the drug until with the patient.
A medication dosage simulation strategy to improve patient safety
A medication dosage simulation project has shown positive outcomes. Medication dosage calculation proficiency is critical to safe nursing care. In most clinical and academic settings, nurses must demonstrate 100% accuracy with medication dosage administration. The time required to develop, administer, grade, remediate, and re-examine a valid ...
MISATO: machine learning dataset of protein-ligand complexes for
MISATO is a database for structure-based drug discovery that combines quantum mechanics data with molecular dynamics simulations on ~20,000 protein-ligand structures. The artificial intelligence ...
A Guide to Your Ozempic Dosage
The recommended starting dose of Ozempic is 0.25 milligrams (mg) once weekly. This low dose will help your body get used to the medication. After four weeks, the dose is typically increased to 0.5 mg once weekly, the maintenance dosage. This article will review how to take Ozempic safely, including commonly recommended dosages, adjusting ...
HESI: Dosage Calculations Assignment Quiz Flashcards
The maximum bolus is 4,000 units and the maximum initial infusion is 1000 units/hr. Determine the client loading dose: 60 units x 90 kg = 5,400 unit. Per protocol, use the maximum loading dose of 4,000 units, and calculate the client's dose using the formula D/H x Q = 4,000 units / 5,000 units x 1 ml = 4/5 = 0.8 ml.
The Assignment with Audie Cornish
6 PM ET: Mass displacement in Gaza, Slovakia PM's condition, drug overdose deaths decline & more. 5 Things. Listen to CNN 5 Things. Wed, May 15 ... The Assignment with Audie Cornish. May 16, 2024.
Dosage Calculations HESI Assignment Exam
Quiz yourself with questions and answers for Dosage Calculations HESI Assignment Exam, so you can be ready for test day. ... The drug is dispensed as a econdary infusion labled, 0.9% normal saline 50 ml with propofol 100 mg/50 ml The client's weight is 198 pounds. The nurse should program the infusion pump to deliver how many ml/hour? (Enter ...
HHS Drops 'Stacking' Provision in Medicaid Drug Rebate Proposal
In an update released Wednesday, the Centers for Medicare & Medicaid Services said it won't finalize a provision in the proposed rule ( RIN 0938-AU28) that would have required companies to "stack" discounts and rebates throughout a transaction when reporting best prices for the Medicaid drug rebate program. Instead, the agency said it ...