crumblr

crumblr enables analysis of count ratio data using precision weighted linear (mixed) models by applying an asymptotic normal approximation of the variance following the centered log ratio transform (CLR) used in compositional data analysis, providing a fast, flexible alternative to GLMs and GLMMs while retaining high power and controlling the false positive rate.

Key Features:

• Precision weighted linear (mixed) models: Enables analysis of count ratio data with precision weighting in linear and linear mixed model frameworks.
• Asymptotic normal variance approximation: Uses an asymptotic normal approximation of the variance following the centered log ratio transform (CLR).
• Compositional data analysis integration: Implements the CLR transform that is widely used in compositional data analysis.
• Alternative to GLMs and GLMMs: Provides a fast, flexible alternative to generalized linear models and generalized linear mixed models.
• Statistical performance: Retains high power and controls the false positive rate.

Scientific Applications:

• Count ratio data analysis: Enables statistical analysis of count ratio data in biomedical and compositional research settings.

Methodology:

crumblr applies an asymptotic normal approximation of the variance following the centered log ratio transform (CLR), then fits precision weighted linear (mixed) models.