#5 · Systems biology

Design-aware statistical models for differential analysis of omics and imaging data

The dominant thread is matching inferential models to the design and structure of omics data rather than to a shared modality. A mass spectrometry imaging workflow provides an open-source R implementation of differential analysis under complex designs, and spatialFDA tests differential cellular co-localisation across multi-sample, multi-condition experiments by combining spatial statistics with functional data analysis. cpam fits shape-constrained changepoint additive models to temporal omics data, including pseudobulked series. A benchmark of precision matrix estimators for differential co-expression networks uses simulated expression data with known ground truth. The GTEx analysis introduces a Gene Stability Score for age- and sex-associated differentially variable genes across 30 tissues, and the DRG nociceptor study builds subtype-resolved protein-protein interaction networks from single-nuclei transcriptomes. What is new is explicit treatment of experimental design, temporal shape constraints, and inter-individual variability as inferential targets. The circadian liver isoform study is less directly aligned with this emphasis.