ExploreModelMatrix: Interactive exploration for improved understanding of design matrices and linear models in R.

Area of Science:

• Statistics
• Bioinformatics
• Computational Biology

Background:

• Linear and generalized linear models are fundamental in scientific research for data analysis and hypothesis testing.
• Interpreting model coefficients and formulating contrasts can be challenging for users, leading to uncertainty.

Purpose of the Study:

• To introduce ExploreModelMatrix, an R/Bioconductor package designed for interactive exploration of linear model design matrices.
• To aid users in understanding model coefficients, formulating contrasts, and performing model diagnostics.

Main Methods:

• The package takes a sample data table and a design formula as input.
• It interactively visualizes how model coefficients combine to produce fitted values.
• It displays diagnostic statistics like variance inflation factors and the pseudoinverse of the design matrix.

Main Results:

• ExploreModelMatrix facilitates the interpretation of individual coefficients within complex linear models.
• Users can easily formulate and test scientific hypotheses through linear contrasts.
• The package provides insights into the properties of the linear model design.

Conclusions:

• ExploreModelMatrix enhances the understanding and application of linear and generalized linear models.
• It serves as a valuable tool for teaching, self-learning, and assisting researchers in interpreting complex model designs.