A modular, mechanism-informed biomarker panel for asthenozoospermia based on integrative sperm proteomics

Asthenozoospermia (AZS), defined by reduced sperm motility, is a major cause of male infertility. Comparative sperm proteomic studies of AZS versus normozoospermic men have reported differential protein expression; however, their biological interpretation remains fragmented and largely centered on individual proteins.

We performed an integrative analysis of 16 independent human sperm proteomic datasets using a two-step recurrence-based strategy to identify robust AZS-associated molecular signatures. Differentially expressed proteins (DEPs) reported in at least two independent studies were considered recurrent and subjected to Gene Ontology, KEGG, Reactome, and Human Phenotype Ontology (HPO) enrichment analyses. The most highly recurrent DEPs within each direction of regulation (upregulated and downregulated) were further selected as core DEPs and analyzed through protein-protein interaction network analysis.

Among 2,459 non-redundant DEPs, 701 proteins (259 upregulated and 442 downregulated in AZS) were recurrently identified across studies, representing approximately 29% of all reported proteins. Recurrently upregulated proteins were primarily associated with proteostasis-related processes, RNA-related pathways, vesicle-mediated trafficking, and extracellular seminal plasma components, whereas recurrently downregulated proteins were linked to mitochondrial energy metabolism and flagellar structure. Protein-protein interaction network analysis of the core DEPs, integrating both upregulated and downregulated sets, revealed four conserved functional modules across sperm compartments: structural, metabolic, vesicular trafficking, and extracellular.

These findings show that molecular alterations in AZS converge at the level of functional modules rather than representing isolated protein-level dysregulation. This modular architecture defines a mechanism-informed biomarker panel that captures key processes underlying AZS, helps reconcile discrepancies among studies, and provides a framework for molecularly grounded diagnostics and future validation through multi-omics and artificial intelligence approaches.