Table of Contents

Graphical Analysis Using the Graphical User Interface

Overview

Users will construct a number of pharmacokinetic graphs using the Graphical User Interface. In addition to creating the plot, users will learn how to do the following:

a) import and transform data

b) modify the plot colors, symbols, text, and graph layout

c) add symbols, Greek letters, curve fit equations, and reference lines

d) subset data for plotting using the menu dialogue or using graphical point selection

e) export graphs to other applications like Word and PowerPoint

f) automating graph production by saving the Script code

S-Plus Graphical User Interface Exercises

1. DATA IMPORT AND TRANSFORMATION

1.1. Import a Multi-Page WinNonlin Output in Excel Format with Missing Values

1.2. Stack Time Concentration Data

1.3. Add Dose to all Records in a NONMEM Input File using Subset and Merge

2. UNIVARIATE PLOTS

2.1. Histogram Density Plot of Clearance

2.2. Review of Graph Hot Spots

2.3. Box-and Whisker Plot of Clearance

2.4. Quantile-Quantile (QQ) Normal Line Plot of Clearance

2.5. Multiple Graphs on a Page

2.5.1. Copy and Paste method

2.5.2. Insert Graph drop-down menu method

2.5.3. Shift-click method

3. BIVARIATE PLOTS

3.1. Grouped Boxplots

3.1.1. Cmax and AUC by Treatment

3.1.2. Cmax by Subject after Multiple Treatments

3.1.3. Cmax by Treatment Conditioned by Sequence

3.1.4. Cmax Conditioned by Sex

3.2. Scatter Plot of Dependent Variable (concentration) vs Weighted Residuals with a Reference line and visually identified extreme value (outlier) labeled

3.3. Scatter Plot of DV vs Weighted Residuals with a Loess Smoother

3.4. Scatter Plot of DV vs Weighted Residuals with a Loess Smoother and Subject ID as Plot Symbols

3.5. Scatter Plot of Time vs Weighted Residuals with a Loess Smoother

3.6. Boxplot of Time vs Weighted Residuals

3.7. Boxplot of ID vs Weighted Residuals

3.8. Visual Identification of Outliers, Linear Fits with and without identified Outliers

3.9. Dotplot of Cmax by Subject Conditioned on Days on Treatment (single and multiple dose)

3.10. Dot Plot of Clearance by Subject, with Symbols Colored by Sex

4. MULTIVARIATE PLOTS

4.1. Correlation Matrix of Random Effects with Loess Smoother

4.2. Correlation Matrix of Random Effects Conditioned by Sex with Loess Smoother

5. ADVANCED GRAPHICS

5.1. Time vs Mean Concentration Error Bar Plot, One Treatment

5.2. Time vs Mean Concentration Error Bar Plot, Multiple Treatments

5.3. Treatment vs PK Parameter Scatter Plot labeled by Subject ID

6. RESPONSE vs TIME PLOTS

6.1. Unstacked Data

6.1.1. Time vs Concentration, Multiple Y Axis, with a Loess Smoother

6.1.2. Time vs Concentration, Multiple Y Axis, with a Loess Smoother Conditioned on Dose

6.1.3. Time vs Concentration, Multiple Y Axis, with a Loess Smoother Conditioned on Subject

6.1.4. Time vs Concentration by Subject, Day, Grouped by Treatment

6.1.5. Time vs Concentration, Population Predicted, and Individual Predicted by Subject

6.2. Stacked Data

6.2.1. Time vs Concentration by Subject, Day, Grouped by Treatment

6.2.2. Time vs Observed and Predicted for Drug and Metabolite by Subject

7. UTILITIES

7.1. Multi-Page Graph Export

7.2. Export to PowerPoint

7.3. Saving Script Files and Automation

S-Plus Applied to Pharmacokinetic Graphical Analysis

S and S-Plus Command Line Programming

Overview

This section focuses exclusively on applying the S programming language to pharmacokinetic graphical analysis. The primary goal is to ensure user’s can construct typical plots used in Study Reports and associated Appendices. The course outline includes a number of introductory exercises designed to take the user through each function independently, before using all the functions in combination to create typical study report plots. Some users may opt not to work through all of the introductory exercises. This section covers the topics listed below.

Fundamentals

Programming Basics
Graphics

Typical Plots
Time vs Response Plots with Emphasis on Overlay Plots

GroupedData Plots
Trellis Graphics
Traditional Graphics

Tree Regression
Population Pharmacokinetic Modeling

S and S-Plus Command Line Programming Exercises

1. Introduction to S Programming

1.1. Subscripting (reference data elements, select and subset data)

1.2. Operators and Functions

1.3. Functions

1.4. Writing Functions

1.5. Conditionals

1.6. Iteration

2. Introduction to Graphics

2.1. General Exercises

2.1.1. simulate x and y vectors to create simple plots used to illustrate functions below

2.1.2. identify() to label points

2.1.3. text() to add labels to points

2.1.4. title() to add a main title

2.1.5. abline() to add a least squares or robust regression line

2.1.6. locator() function to interactively add a legend

2.1.7. commands to save and restore original par() settings

2.1.8. pch and col to change colors and symbols

2.1.9. key() to add a legend

2.1.10. utilities (appendix)

2.2. Fundamentals Applied to Pharmacokinetic Plots

2.2.1. Adding data variables as plot symbols

2.2.2. Adding lines to a plot

2.2.3. Adding text to plots

2.2.4. Adding lines to plots for different groups and data variables as plot symbols

2.2.5. Adding text to margins

3. Univariate Plots

3.1. Histograms and multiple graphs on a page

3.2. Density plot

3.3. Boxplots

4. Bivariate Plots

4.1. Boxplots by treatment group

4.2. Trellis histogram by treatment group

4.3. Trellis density plot by treatment group

4.4. Trellis overlay dotplot by treatment

4.5. Trellis boxplot by treatment

4.6. Trellis strip plot with jitter by treatment

4.7. Error bar plot by treatment

5. Multivariate Plots

5.1. Pairs plot with a loess smoother

6. Time vs Response Plots

6.1. Grouped Data

6.1.1. Time vs concentration by subject

6.1.2. Time vs concentration subject overlay

6.1.3. Time vs concentration subject overlay by dose

6.1.4. Time vs concentration per PK compartment by subject

6.1.5. DeltaBP vs dose following multiple treatments by rabbit (crossover study)

6.1.6. Concentration vs uptake rate by location*treatment

6.1.7. Age vs Metric subject overlay by sex

6.2. Trellis Overlay Plots using par(new=T)

6.2.1. Time vs concentration and time vs predicted of drug and metabolite by subject

6.2.2. Time vs concentration following multiple treatments by subject and day (single and multiple dose crossover trial)

6.3. Trellis Plots and Panel.Superpose

6.3.1. Time vs concentration and time vs predicted of drug and metabolite by subject

6.3.2. Time vs concentration and time vs population predicted and time vs individual predicted with the subset option by subject

6.4. Traditional Graphics

6.4.1. Function to plot time vs concentration and time vs log(concentration) following multiple treatments by subject and with the option to specify different page formats

6.4.2. Function to plot time vs concentration with multiple treatment overlay by subject and with the option to specify different page formats

6.4.3. Function to plot time vs concentration, time vs predicted, multiple treatments overlay by subject and with the option to specify different page formats

7. Tree Regression

7.1. Grow and display a tree using Cl as the dependent variable and a set of patient demographic and potential baseline covariates

7.2. Prune a tree

8. Population PK Modeling

8.1. Algorithms – nls(), nlslist(), nlme()

8.2. Built-in Self-Start Models

8.3. Writing Models in S

8.4. Writing Models in C

9. Project Management

9.1. Set-up different S-Plus databases for management of objects by project