The authors examined 65 full-length SRY proteins and 38 HMG box subsequences from 75 mammalian species using alignment-independent, quantitative measures. They did not rely on sequence alignment to infer similarity; instead, they computed amino-acid and dipeptide compositions, Shannon entropy, dominant-residue profiles, hydropathy-based n-gram spectra, the longest hydropathy stretches, and cumulative proximity. The HMG box emerged as the more compositionally constrained unit. Basic residues predominated in HMG domains, cysteine was nearly absent, and entropy values clustered near 4.0 in the HMG set. The authors reported an average polar residue content of 58.44% for HMG boxes, and noted that 25 of 38 HMG domains shared a polar-residue fraction of roughly 57.97%.
Full-length SRY proteins showed greater compositional heterogeneity. Rodent sequences trended toward polar, glutamine-rich profiles. Cetacean sequences exhibited lower polar content. Terminal regions (the divergent N- and C-termini) lacked sequence homology by alignment, yet the analysis revealed that these regions carry compositional signals that differ from those of the HMG box; they are not simply passive linkers, according to the data.
A multifactorial approach underpinned the comparative results. Amino-acid composition and dipeptide frequencies captured residue-level biases. Shannon entropy provided a compact metric of compositional complexity. Hydropathy n-gram spectra and the measurement of longest hydropathy stretches added information about contiguous hydrophobic or hydrophilic runs.
Data indicated the HMG box retains a narrow compositional grammar: basic residues dominate and cysteine is nearly absent. The authors reported entropy clustering near 4.0 for HMG domains, consistent with constrained compositional variability. By contrast, full-length SRY profiles showed wider dispersion across these same measures. This divergence in dispersion produced differential clustering patterns when proximity metrics were applied.
Across the 65 SRY proteins and 38 HMG domains, predicted structural state was dominated by intrinsic disorder. The study team observed that intrinsically disordered predictions were pervasive, reflecting the compositional properties of both terminal regions and many full-length proteins. They also documented a specific caveat: a polyglutamine (polyQ)–rich rodent SRY exemplified how low-complexity repeats can yield misleading predictions that favor ordered structure under some algorithms. The authors therefore recommended consensus disorder analysis when interpreting rapidly evolving, compositionally biased regions, because single-method predictions may be confounded by low-complexity tracts.
The source did not provide experimental structural validation or specify which disorder predictors were used in every comparison. Those methodological details were not reported in the abstract.
The investigators applied cumulative proximity analysis as a multivariate distance metric. This approach yielded taxonomically consistent clustering at the HMG box level more often than at the full-length SRY level, where terminal-region variation blurred domain-level signals. The study identified Bos taurus and Bos grunniens as the only species pair that clustered proximally at both the full-length SRY and the HMG-box evaluations, suggesting that meeting multivariate proximity thresholds across both scales is a stringent criterion of compositional similarity.
Proximity-based clustering therefore appears to be more informative for conserved domains than for variable terminal regions. The authors used this observation to argue that compositional individuality is common across mammalian SRY sequences: alignment-independent signatures revealed distinct local grammars rather than a single, uniform sequence composition.
The authors proposed that quantitative compositional signatures could provide a framework for interpreting SRY variation and assessing clinical mutations, by supplying alignment-independent metrics to contextualize sequence changes. They argued that compositional individuality—observed at both full-length and HMG-box levels—should inform how one evaluates the potential functional impact of variants, particularly in regions where alignment-based homology is uninformative.
The preprint explicitly noted limitations and next steps. Denser taxon sampling is needed to refine compositional boundaries across mammals. Structural modeling and experimental tests remain necessary to determine whether the observed HMG-box drift and terminal-region variation translate into altered DNA binding, bending, or reproductive function. The abstract did not report details on experimental validation, specific statistical thresholds, or exact computational tools used for every analysis; those methodological specifics were not provided in the text that served as this summary. These gaps frame the immediate follow-up work: validate sequence-derived compositional signals with structure and function assays, expand species coverage, and apply consensus disorder pipelines where low-complexity repeats are present.