BMDExpress 3.20.0106

New Features

• Added venn diagram visualizations to category analysis and BMD analysis results viewers.
• Add upstream analysis notes to exported files.

Bug Fixes

• Fixed an issue flagging hill models when BMD is below the lowest dose. Previously the application interpreted 1/3 as 1/4 and 1/2 as 1/3 due to a bug creating the flag hill ratio value.

BMDExpress 3.20.0095

New Features

• Add Adverse Direction (dose-response curve direction of response) filter to Category Analysis to allow for gene set analysis of up- or down-regulated genes in isolation.
• Update labels in Category Analysis input view to be consistent and explicit
• To help with empirical false discovery analysis the total possible geneset counts for a data set based on platform annotation and gene set size filter were added to the Category Analysis notes.

Bug Fixes

• Fixed issue causing Polynomial models to produce non-monotonic results when user specified monotonic restriction

BMDExpress 3.20

New Features

• R-squared (coefficient of determination) as an alternative to global goodness of fit p-value for evaluating fit of models to data
• Step Function detection with user defined thresholds for application to modeled curves to facilitate identification and removal of biologically implausible responses
• Export fitted curve response data
• Add category analysis filter parameters
  • Predefine the Min/Max number of genes in gene set
  • Remove fits that are step functions
  • Remove fits that are step functions with BMD less than lowest dose
  • Filter fitted responses based on R-squared

Bug Fixes

• Fixed issue when running multiple One Way ANOVA analyses causing LOTEL/NOTEL
  calculation to shuffle p-value and fold change parameters in GUI. This was not an issue when running one dataset at a time and in the command line version
• Fixed issue cause by invalid default folder path for exporting and importing data.

BMDExpress 3.1

• Enable relative deviation during BMD Analysis for non-log transformed data
• QC Bar chart for expression data to show gene expression direction per dose based on T-Test
• Polynomial 2 model now has option to be non-monotonic
• Add polynomial 2 model to curve fit prefilter
• Add ability to save curve fit prefilter settings

BMDExpress 3

• Laplace and MCMC Bayesian model averaging (based the approach used in ToxicR package -- https://github.com/NIEHS/ToxicR)
• Curve fit-based prefilter to identify dose responsive features more effectively
• New maximum likelihood estimate models currently employed EPA's BMDS software for best model-based approaches
• Alternative confidence interval estimate method (Wald) that allows for more rapid curve fitting
• Forward toxicokinetic modeling to estimate internal dose levels based upon external dose BMD value