Finch Studio

# Interactive Plots and Tables

Population Plot Individual Plots Covariate Distribution Plot Individual Fit Plots Residual Diagnostics Plots Residual Distribution Plot Parameter Distribution Plot Covariate vs. Parameter Plots

Finch Studio provides a suite of interactive plots and tables that are seamlessly available to allow the modeler to quickly explore their data, understand their modeling results, triage and debug model or dataset issues, and make decisions on next steps in their analyses. Options are present to color, stratify, or filter by covariates of interest, show individual subject dosing times, add dataset variables to the interactive tooltip, and more.

# Requirements

# Observed Data Plots and Tables

Plot/Table Description Requirements
Individual Plots Each subject in the dataset is shown on a seperate plot. By default TIME vs DV columns are plotted.
  • NONMEM formatted dataset with, at a minimum, ID, TIME, DV, and EVID columns.
  • Covariates for stratifying/coloring/filtering are determined from data mapping.
Population Plots Spaghetti plot of all subjects in the dataset. Hover over lines to see subject ID. By default TIME vs DV columns are plotted
  • NONMEM formatted dataset with, at a minimum, ID, TIME, DV, and EVID columns.
  • Covariates for stratifying/coloring/filtering are determined from data mapping.
Covariate Distribution Plot Matrix of plots showing the distribution of selected covariates in your input dataset on the diagonals, and on the off-diagonals, correlation plots of the selected covariates are shown. NONMEM formatted dataset. Mapped covariate columns will be available to select.
Observed Summary Stats Table Summary statistics of your input dataset. NONMEM formatted dataset.

# Results Plots and Tables

The modeling results plots and tables (individual fits, diagnostic plots, parameter distribution plots, etc) are reliant on the correct NONMEM output tables being present after the model finishes running. Naming conventions for NONMEM output tables required by Finch Studio follow xpose conventions. The key tables that are needed are detailed below. The file names of the tables should always be named exactly as shown (sdtab1, patab1, cotab1, and catab1), since that is what Finch Studio looks for to create the interactive plots. Do not change the number at the end of the table names, this should always be 1.

Table Description Example
sdtab1
  • Standard table file, containing ID, TIME,, MDV, EVID, DV, PRED, IPRED, WRES, IWRES, RES, IRES, etc.
  • Place any variable that can vary within a subject in sdtab1 (TSLD, PERIOD, BLQ, etc.)
$TABLE ID TIME TSLD DV MDV EVID IPRED BLQ CWRES IWRES ONEHEADER NOPRINT FILE=sdtab1
patab1
  • Parameter table, containing individual and/or typical value model parameters
  • ETAS(1:LAST) will include all ETA parameters
 $TABLE ID KA CL V ETAS(1:LAST) ONEHEADER NOPRINT FILE=patab1 ; model parameters
catab1 Categorical covariates (e.g. SEX, RACE, etc) $TABLE ID RACEN ONEHEADER NOPRINT FILE=catab1 ; categorical covariates
cotab1 Continuous covariates (e.g. WTKG, AGE, etc) $TABLE ID WTKG ONEHEADER NOPRINT FILE=cotab1 ; continuous covariates

Example code in NONMEM control stream to request the output tables required by Finch Studio:

$TABLE ID TIME DV MDV EVID IPRED CWRES IWRES ONEHEADER NOPRINT FILE=sdtab1
$TABLE ID KA CL V ETAS(1:LAST) ONEHEADER NOPRINT FILE=patab1 ; model parameters
$TABLE ID RACEN  ONEHEADER NOPRINT FILE=catab1 ; categorical covariates
$TABLE ID WTKG   ONEHEADER NOPRINT FILE=cotab1 ; continuous covariates
Plot/Table Description Requirements
Individual Fits Individual subject and population predictions overlayed on individual's observed data. Solid lines represent individual prediction, dashed lines represent population prediction, and points represent observed data. NONMEM Output Tables: sdtab1, patab1, cotab1, catab1
Residuals vs Time, DV Four panel plot of CWRES vs PRED, CWRES vs TIME, DV vs PRED, and DV vs IPRED. Options are available to change the residual and/or independent varaibles (i.e. TIME, TSLD) being plotted in the plot settings menu. NONMEM Output Tables: sdtab1, patab1, cotab1, catab1
Residual Distribution Distribution (density and histrogram overlays) of residuals. NONMEM Output Tables: sdtab1, patab1, cotab1, catab1
Covariates vs Parameters Matrix of plots of selected covariates vs selected parameters NONMEM Output Tables: sdtab1, patab1, cotab1, catab1
Parameter Distribution Matrix of plots showing the distribution of selected parameters on the diagonals, and on the off-diagonals, correlation plots of the selected parameters are shown. NONMEM Output Tables: sdtab1, patab1, cotab1, catab1
Parameter/OFV Convergence OFV and parameter search history from the model run. Can be updated live as the model is running to track progress. .ext file within the model directory, or within the most recently created modelfit_dir subdirectory.
Individual OFV Contribution Each individual's contribtuion to the OFV, sorted from largest to smallest.

mod1.phi file.

Note, if mod1.file is not present after the run completes, please ensure the PsN execute run command is specifying the .phi file to be retained (e.g. execute mod1.ctl -nm_output=xml,ext,phi)

Parameter Correlation This plot shows the strength and direction of the linear relationship between each pair of parameters. Darker color represents stronger correlation. mod1.xml file.
Results Summary Stats Summary statistics of results tables. For example, summary statistics for drug clearance (or any other parameter) across cohorts, race, or any other stratifying variable of interest can quickly be generated NONMEM Output Tables: sdtab1, patab1, cotab1, catab1
SCM Results Summary of stepwise-covariate-modeling run, showing which covariate was added or removed at each step. scmlog.txt file within an scm_dir directory.

# Stratifying, Coloring, and Filtering Plots

All plots within Finch provide you the ability to stratify, color (group), and filter your data to give you the flexibility you need to fully explore and understand your data. When applicable, you may also choose the X and Y variables to plot.

# Plot Settings

# Grid Layout

Specify the number of rows and columns for displaying plots. (i.e. 3 rows and 3 columns will display up to 9 total plots)

# X-axis and Y-axis Settings

  • Toggle between linear or semi-log axes

  • Choose whether the axes should be free (autoscales based on the data within each plot) or fixed (consistent across plots).

  • Axis Title

# Plot Options (Layers)

Different plots will allow you to toggle different layers. For example, for the population plot you can show lines, points, and mean lines.

# Tooltip Variables

When hovering over a point, Finch will display the subject ID number. If you wish to see more information about a point, you may add additonal variables to the tool tip.

# Plot Types - Observed Data

# Population Plots (Spaghetti Plot)

An image

# Individual Plots

An image

# Covariate Distribution (and Correlations)

An image

# Plot Types - Model Results

# Individual Fits

An image

# Model Diagnostics

# RES vs TIME, RES vs DV, IPRED vs DV, PRED vs DV

An image

# Residuals Distribution

An image

# Parameter Distribution

An image

# Covariate vs Parameter

An image

← Changelog and Release Notes →