Genomic conservation shapes predictability of the human gut resistome from microbial community composition

This dataset contains the processed data products and analysis outputs supporting the study “Genomic conservation shapes predictability of the human gut resistome from microbial community composition.” The deposit includes derived taxonomic and resistome matrices, model prediction outputs, pangenome and genomic conservation evidence, provenance information, and the result tables underlying the main and supplementary analyses. The files are organised as follows. matrices.zip contains the processed and derived matrices used in the downstream analyses. model_outputs.zip contains prediction outputs from cross-validation, leave-one-study-out evaluation, subject-grouped validation, and held-out cohort analyses. pangenome.zip contains the genomic evidence used to estimate within-species conservation and host breadth, including the information supporting the conservation–predictability analyses. tables.zip contains the result tables supporting the main and supplementary analyses. provenance.zip contains provenance and supporting information used to document the construction and origin of the derived datasets. Zenodo_README.md describes the deposited files, their organisation, and their relationship to the analysis workflow. The study examines the relationship between microbial community composition and the antimicrobial resistome across human gut and other metagenomic habitats. It further evaluates cross-cohort prediction of antimicrobial resistance genes and tests whether variation in gene-level predictability is associated with within-species genomic conservation. Raw resistome alignment data, taxonomic profiles, metadata, and reference genome resources are available from their original public repositories and are not redistributed in this deposit. Analysis code is maintained separately on GitHub.