Computer-aided formulation development MCQs With Answer

Introduction:
This set of multiple-choice questions focuses on Computer-Aided Formulation Development, a critical area in the M.Pharm Computer Aided Drug Development (MPH 203T) syllabus. The questions cover experimental design, predictive modeling, in silico tools for excipient selection and stability prediction, process simulation, PAT integration, multivariate analysis, and regulatory considerations linked to Quality by Design (QbD). They are designed to deepen conceptual understanding and application skills, preparing students for formulation design, optimization, and troubleshooting using computational approaches. Each question emphasizes practical relevance, model selection, interpretation of outputs, and integration of in silico methods into laboratory and scale-up workflows.

Q1. Which experimental design is most efficient for fitting a quadratic response surface when three factors are being optimized?

• Full factorial design (2^3)
• Box-Behnken design
• One-factor-at-a-time design
• Plackett-Burman design

Correct Answer: Box-Behnken design

Q2. In Quality by Design (QbD) for formulation development, the term “design space” refers to which of the following?

• The fixed set of process parameters used during routine production
• The multidimensional combination of input variables and process parameters that assure product quality
• The list of raw materials approved for use in a product
• The standard operating procedures for the analytical lab

Correct Answer: The multidimensional combination of input variables and process parameters that assure product quality

Q3. Which computational method is most appropriate for predicting excipient compatibility with a new API based on molecular interactions?

• Principal component analysis (PCA)
• Molecular docking and molecular dynamics simulations
• Plackett-Burman screening
• Linear regression of dissolution data

Correct Answer: Molecular docking and molecular dynamics simulations

Q4. In DoE analysis, which statistical output primarily indicates whether a factor has a statistically significant effect on response?

• R-squared value
• p-value from ANOVA
• Standard deviation of the factor
• VIF (variance inflation factor)

Correct Answer: p-value from ANOVA

Q5. Which modelling approach is best suited for capturing complex nonlinear relationships between formulation variables and tablet tensile strength?

• Multiple linear regression
• Artificial neural networks
• Simple moving average
• Univariate t-test

Correct Answer: Artificial neural networks

Q6. What is the primary aim of physiologically based pharmacokinetic (PBPK) modelling in formulation development?

• To predict mechanical properties of tablets
• To simulate drug absorption, distribution, metabolism and excretion using physiological parameters
• To compute dissolution media composition
• To design coating equipment settings

Correct Answer: To simulate drug absorption, distribution, metabolism and excretion using physiological parameters

Q7. Which descriptor type is commonly used in QSAR models to relate chemical structure to formulation-relevant properties?

• Process flow diagrams
• Molecular descriptors (e.g., logP, molecular weight, polar surface area)
• Manufacturing batch records
• Instrument calibration curves

Correct Answer: Molecular descriptors (e.g., logP, molecular weight, polar surface area)

Q8. In multivariate data analysis for PAT, which technique is widely used for monitoring batch-to-batch variability and detecting deviations online?

• Hierarchical clustering
• Principal component analysis (PCA) and multivariate control charts
• Unpaired t-test
• Kaplan-Meier survival analysis

Correct Answer: Principal component analysis (PCA) and multivariate control charts

Q9. When building a predictive dissolution model, which input is least likely to improve model generalizability?

• Physicochemical properties of API and excipients
• Experimental dissolution profiles under varied conditions
• High collinearity redundant features without regularization
• Process parameters like compression force and granulation speed

Correct Answer: High collinearity redundant features without regularization

Q10. Which in silico tool would you choose to predict long-term chemical stability (degradation pathways) of an API in a formulation?

• Design of Experiments (DoE)
• Reaction pathway prediction and accelerated stability modelling using kinetic models
• Hot-melt extrusion simulation
• Tablet hardness tester

Correct Answer: Reaction pathway prediction and accelerated stability modelling using kinetic models

Q11. Box-Behnken and Central Composite Designs differ mainly because:

• Box-Behnken is for qualitative responses; CCD is for quantitative responses
• CCD includes axial (star) points to estimate curvature while Box-Behnken does not use extreme corner points
• Box-Behnken requires fewer runs than any fractional factorial design
• CCD cannot model quadratic terms

Correct Answer: CCD includes axial (star) points to estimate curvature while Box-Behnken does not use extreme corner points

Q12. Virtual formulation screening primarily helps to:

• Eliminate the need for any lab experiments
• Prioritize excipient/API combinations and reduce experimental workload
• Replace regulatory submission requirements
• Directly control manufacturing line speed

Correct Answer: Prioritize excipient/API combinations and reduce experimental workload

Q13. Which validation metric is most informative for checking predictive performance of a regression model in formulation optimisation?

• Confusion matrix
• Root mean square error (RMSE) on an independent test set
• Chi-square goodness of fit test
• Log-rank statistic

Correct Answer: Root mean square error (RMSE) on an independent test set

Q14. Which method is appropriate for feature selection when many correlated descriptors are available for a QSAR model?

• Leave all descriptors unchanged
• Use regularization techniques such as LASSO or elastic net
• Only use descriptors with the highest mean values
• Randomly drop descriptors until model fits training data

Correct Answer: Use regularization techniques such as LASSO or elastic net

Q15. In-silico IVIVC (in vitro–in vivo correlation) modelling is useful because it can:

• Predict which manufacturing operator will produce the best tablets
• Allow prediction of in vivo performance from in vitro dissolution profiles to support formulation decisions
• Ensure zero variation between batches
• Determine the chemical structure of unknown impurities

Correct Answer: Allow prediction of in vivo performance from in vitro dissolution profiles to support formulation decisions

Q16. Which of the following is a primary advantage of using mechanistic models (e.g., heat and mass transfer) in process scale-up simulations?

• They eliminate the need for pilot-scale runs entirely
• They provide insight into physical phenomena enabling rational scale-up and equipment selection
• They are always simpler than empirical models
• They require no physical property inputs

Correct Answer: They provide insight into physical phenomena enabling rational scale-up and equipment selection

Q17. Which approach best addresses overfitting when developing machine learning models for formulation prediction with limited experimental data?

• Use very deep neural networks without regularization
• Apply cross-validation, simpler models, and regularization
• Only evaluate model performance on the training set
• Increase the number of features without increasing samples

Correct Answer: Apply cross-validation, simpler models, and regularization

Q18. Which PAT (Process Analytical Technology) tool is commonly integrated with multivariate models to monitor granulation endpoint in real time?

• NMR imaging of final tablets
• Near-infrared (NIR) spectroscopy combined with multivariate calibration
• Visual inspection by an operator only
• High performance liquid chromatography on-line for real-time moisture

Correct Answer: Near-infrared (NIR) spectroscopy combined with multivariate calibration

Q19. Which statement about cheminformatics fingerprinting used in formulation research is correct?

• Fingerprints are high-dimensional binary vectors encoding presence or absence of substructures
• Fingerprints measure tablet hardness directly
• Fingerprints are synonymous with dissolution profiles
• Fingerprints are used to set compression force on a press

Correct Answer: Fingerprints are high-dimensional binary vectors encoding presence or absence of substructures

Q20. Which regulatory concept supports use of in silico models and simulation for reducing experimental burden during formulation development?

• Use of One-Factor-At-A-Time (OFAT) approach
• Quality by Design (QbD) and model-informed drug development (MIDD)
• Complete elimination of stability studies
• Mandatory full-scale production without pilot studies

Correct Answer: Quality by Design (QbD) and model-informed drug development (MIDD)

Download