#38 · Genomics

Normalizing sequence-derived scores against length, alphabet, and composition effects

The common thread is making the components of sequence-derived measures explicit so that values can be corrected or inspected before interpretation. The Effort-To-Compress preprint proposes normalized variants, nETC and cETC, with certified bounds and matched references across sequence length, alphabet size, and composition. The KmerSignificance score (KSS) combines an information-theoretic discriminative measure, a GA-optimized MIYATA_EVO substitution matrix, and a UniProt-based component to prioritize viral k-mers. The mammalian SRY study is more descriptive, applying alignment-free compositional analysis, disorder, and proximity clustering to 65 SRY proteins and 38 HMG boxes across 75 species. What is new is explicit quantitative correction of scores whose raw values would otherwise mislead, rather than single-valued outputs taken at face value.