New Mathematical Method for Analyzing Hierarchical Compositional Data in Biology
Researchers have developed PolyILR, a new mathematical decomposition method for analyzing compositional data (vectors representing relative proportions) that respects both hierarchical structure and geometric properties. The method addresses limitations in existing approaches by preserving the Aitchison geometry while aligning with tree-structured data like taxonomies and phylogenies. This work is significant for fields like microbiome analysis and genomics where data naturally have hierarchical organization.
PolyILR is a canonical orthonormal decomposition technique designed to handle compositional data—vectors encoding relative proportions—that arise across scientific domains including ecology, geochemistry, and genomics. The key innovation is that it preserves the intrinsic Aitchison geometry while respecting known hierarchical structures such as taxonomies, phylogenies, and ontologies. The method constructs a weighted local geometry at each internal node of a tree to capture full branching structure, then lifts these to a global orthonormal basis where each coordinate corresponds to a specific tree location. Benchmarks on microbiome and single-cell data demonstrate that PolyILR produces stable, interpretable features and enables inference at multiple tree resolution levels. The authors also establish a theoretical connection to softmax classifiers, suggesting potential applications in probabilistic modeling.
What's missing
The paper does not discuss computational complexity or scalability to very large trees, nor does it provide detailed comparisons with specific alternative methods for handling hierarchical compositional data.
What different sources said
- arXiv cs.LGCenter
Tree-Structured Orthonormal Decomposition of the Aitchison Simplex
Related
Genetic Drift, Not Selection, Drives Rapid Feather Color Evolution in Island Bird Radiation
A new study of an island bird radiation found that rapid evolution of feather coloration is driven primarily by genetic drift in small populations rather than sexual or ecological selection. The research integrated whole-genome data with detailed plumage measurements across complete species sampling to test whether signaling trait evolution correlates with speciation rates. The findings suggest that neutral demographic processes play a central role in generating phenotypic diversity during island radiations, challenging assumptions about the mechanisms driving rapid evolution.
New AI Model Improves Prediction of Therapeutic Peptide Function from Protein Sequences
Researchers developed a lightweight CNN classifier that predicts whether peptide sequences have therapeutic properties, trained on a database of 54,655 peptides across 48 functional categories. The model uses a novel negative sampling strategy to reduce false positive rates from over 60% in previous approaches to 2.1%. This advancement could accelerate drug discovery by enabling faster computational screening of peptide candidates before expensive experimental testing.
Study Shows Different Metabolic Stress Models Produce Distinct Effects on Human Neuronal Networks
Researchers tested three common in vitro metabolic stress models on human-derived neuronal networks and found each produced different patterns of neuronal activity and cell damage. The models tested were hypoxia alone, oxygen-glucose deprivation (OGD), and hypoxia combined with glutamate exposure. The findings suggest that choice of experimental model significantly affects results and that combining electrophysiological and structural analyses is important for accurately assessing metabolic stress in stroke research.