#27 · Genomics

Structure, composition and gene order used to infer prokaryotic gene function

The common thread is functional and taxonomic inference that bypasses direct sequence homology, drawing instead on structural domain assignment, k-mer composition and genomic neighbourhood. The archaeal proteome preprint classifies 124,075 proteins at domain level with ECOD, using AFDB and AlphaFold3 structures, and reports a conserved fold repertoire. A second preprint infers genetic codes from assembled genomes with a k-mer method described as more than 100-fold faster, applied to thousands of archaeal and bacterial assemblies. The published [NiFe]-CODH study applies synteny analysis to 1376 CODH and 1545 HCP sequences, linking CODH clades to operon context and catalytic activity. What is new is treating structure, composition and gene order as bases for functional claims. The evidence remains largely preprint, and the CODH link rests on genomic context rather than direct homology.