Molecular signatures of sperm-level selection in humans and their evolutionary conservation

Of the millions of sperm in a human ejaculate, only one fertilises an egg. Selection on sperm is strong, offering an opportunity to proofread novel germline genotypes generated by segregation, recombination, and de novo mutations. Here, we combine within-ejaculate selection on phenotypic sperm quality with deep-coverage whole-genome sequencing and tandem mass tag proteomics to identify molecular biomarkers of human sperm quality. High-quality sperm are characterised by increased longevity, motility, morphological normality, and DNA condensation. We show that high-quality sperm are genetically distinct from low-quality sperm at >1,000 loci mapping to 511 genetic elements, 64% of which fall within protein-coding portions including intronic regions. Genes diverging between high- and low-quality sperm pools are involved in cell signalling, cell proliferation, and neurodevelopmental processes. Polygenic scores differ between high- and low-quality sperm and reveal variants linked to overall health and healthy lifestyle choices, linked by putative negative correlation. TMT proteomics show significant differences in the abundance of 1,178 proteins, with known markers of inflammation, immune function, and cancer-related cell proliferation showing lower abundance in high-quality sperm. The genomic signature of sperm selection in humans and zebrafish shows significant overlap, indicating evolutionarily conserved processes determining sperm function across taxa. Our finding of a direct link between phenotypic variation and genomic variation in human sperm highlights the importance of within-ejaculate selection and its key contribution to determining offspring fitness and health.