Original title: StableMate: a statistical method to select stable predictors in omics data
Authors: Yidi Deng,Jiadong Mao,Jarny Choi,Kim-Anh Lê Cao
The article discusses the importance of identifying statistical associations between biological variables in order to understand molecular mechanisms. However, many association studies based on correlation or linear regression analyses often lack reproducibility and interpretability due to the complexity and variability of omics datasets.
To address these challenges, the researchers developed a regression framework called StableMate. This framework utilizes variable selection across heterogeneous datasets to select predictors that are stable across different environments, as well as environment-specific predictors. Using these predictors, StableMate can build regression models that can make generalized predictions in unseen environments.
The researchers applied StableMate to three case studies: breast cancer RNA-seq data, metagenomics data for colon cancer, and scRNA-seq data for glioblastoma. In each case, StableMate successfully identified genes or microbial signatures that consistently predicted certain characteristics, regardless of disease status or cell location.
Overall, StableMate is a flexible framework that can be used for regression and classification analyses, allowing for comprehensive characterization of biological systems across different omics data types.
Original article: https://www.biorxiv.org/content/10.1101/2023.09.26.559658v2