Comparative biomimetic microfluidics uncovers species-specific sperm navigation across five mammals

Conventional sperm selection for assisted reproduction remains largely divorced from the physical and selective pressures encountered in the female reproductive tract. Here we report, to the best of our knowledge, the first head-to-head, cross-species comparison of microfluidic thigmotactic sperm selection versus density gradient centrifugation (DGC) across five mammalian species, ram, bull, dog, boar and human, using 20 independent ejaculates per species processed in parallel on a single multi-compartment device. Each ejaculate was split between an unprocessed control (RAW), DGC and microfluidic (MF) selection, and the paired data were analysed with mixed-effects models including male as a random effect with ejaculate nested within male, a formally tested species × processing-method interaction, and control of multiplicity across outcomes. Microfluidic selection improved DNA fragmentation, total and progressive motility, acrosome integrity and morphology relative to DGC in every species (all adjusted p ≤ 0.044). The DNA fragmentation index fell from a mean of 19.3% to 2.0% in human sperm and from 11.1% to 1.1% in ram sperm; relative to DGC, the odds of fragmentation were reduced 2.8-fold (95% CI 2.5-3.2) in human and 3.7-fold (3.1-4.5) in ram sperm. The magnitude of the effect differed between species for most outcomes (interaction p < 0.001), being smallest in boar sperm. Changes in kinematic indices were species-dependent and frequently not significant, so the advantage of microfluidic selection is established for the principal quality parameters rather than for every measured variable. These gains reflect enrichment of a competent subpopulation rather than protection of sperm DNA, and were accompanied by a 30- to 50-fold reduction in total motile sperm recovered, an important functional constraint. Navigation success at the most acute (135°) angular challenge was higher for ram sperm than for every other species (differences of 18.4-25.0 percentage points, all adjusted p ≤ 0.004), whereas bull, dog and boar sperm were indistinguishable from one another. Because semen source, cryopreservation history and donor characteristics were not equivalent across species, these interspecies differences are reported as descriptive comparisons and not as evidence of evolutionary adaptation or of biological superiority of any species. This work provides a standardised in vitro framework for comparing sperm selection across mammalian species. Studies reporting fertilisation, embryo development and pregnancy outcomes are required before the platform can be considered for routine assisted reproduction.