This study analyzed human positively selected genes (PSGs) and positively selected residues (PSRs) using a manually curated, high-quality collection of PSGs and PSRs together with a large-scale PSR dataset constrained to proteins with experimental structures. The aim was to clarify residue-level structural principles of adaptive evolution in humans and resolve conflicting conclusions from earlier work, particularly regarding the role of intrinsically disordered regions.
The authors report a consistent picture across datasets that links positive selection to specific protein classes, subcellular localizations, structural features, and host–pathogen interaction interfaces. All findings reported below are drawn from the analysis and integration performed by the authors on these curated and large-scale datasets.
PSRs were significantly enriched in proteins that are secreted or localized to the cell membrane. Within such proteins, PSRs preferentially localized to extracellular regions rather than intracellular compartments. This localization bias implicates external-facing protein surfaces as common targets of adaptive amino acid changes in humans.
The concentration of PSRs in secreted and membrane proteins suggests that evolutionary pressures acting on extracellular encounters—such as interactions with other cells, molecules, or pathogens—contribute substantially to adaptive substitutions observed at the residue level.
At the structural level, PSRs were preferentially found on solvent-exposed surfaces. Secondary structure analysis showed an enrichment of PSRs in coil regions rather than in helix or sheet elements. Importantly, PSRs were also overrepresented at protein–protein interaction interfaces, marking molecular recognition surfaces as hotspots of adaptive change.
Conversely, neither annotated protein domains nor intrinsically disordered regions (IDRs) exhibited enrichment for PSRs in the datasets analyzed. This finding refines earlier, conflicting reports by indicating that structured, solvent-accessible interaction surfaces—particularly extracellular ones—are the primary structural contexts for positive selection, rather than disorder per se or domain boundaries.
Spatial clustering analysis demonstrated that PSRs often accumulate within localized surface patches on protein structures. These clusters indicate that adaptive substitutions are not randomly distributed across exposed surfaces but tend to delineate discrete molecular recognition modules. Such surface patches likely correspond to functional interaction sites that experience recurrent selective pressures.
When PSRs and PSGs were integrated with host–pathogen interaction datasets, a clear pattern emerged: adaptive changes preferentially affect extracellular molecular recognition surfaces and recur in certain pathogen-interacting membrane protein families and particular domain types. Additionally, PSRs showed enrichment in residues that are in direct contact with pathogenic proteins at human–pathogen interaction interfaces.
This enrichment implies that neutralizing pathogen attacks is one of the driving forces behind adaptive evolution of human proteins, with direct-contact residues at interfaces being frequent targets of positive selection.
Analysis of human variation data indicated that PSRs are enriched in substitutions classified as benign and depleted in substitutions classified as pathogenic. This observation extends prior findings that PSGs contain elevated levels of missense variation and suggests that many adaptive residues tolerate variation observed in contemporary human populations, while being less likely to coincide with clinically pathogenic changes.
The authors observed that PSGs in the large-scale dataset were enriched among clinical-stage drug targets. This enrichment suggests a potential link between positive selection and pharmacological relevance: proteins under adaptive pressure, particularly those with exposed extracellular recognition surfaces, may also be frequent targets of therapeutic development or have properties that make them amenable to drug targeting.
Together, these results support a refined structural model of adaptive evolution in human proteins. Specifically, extracellular exposed molecular recognition surfaces and human–pathogen interaction interfaces emerge as recurrent hotspots of positive selection. PSRs form spatially localized surface patches that often correspond to direct-contact residues in host–pathogen interfaces.
The authors propose that precise identification and mapping of human PSRs could help delineate important host–pathogen interfaces that have exerted evolutionary selection pressure and could highlight novel therapeutic targets. Details such as dataset composition, specific protein families implicated, and numerical enrichment values are reported in the original manuscript and supplementary materials; those dataset specifics are not restated here beyond the high-level findings summarized from the source.