Asking the pikas politely…
Microsatellites, also called simple sequence repeats (SSRs), form an ordered repeat-count allele space, yet genetic distances emphasize different distributional features. We developed a multilocus Wasserstein family in which order p varies sensitivity to large repeat displacements while the allele-frequency object, loci, and aggregation remain fixed. Rooted D_p is a metric for real p≥1 on a common locus set; its companion moment D_p^p is generally nonmetric for p>1. Across 300 known-history evaluations, D_1 had a median Spearman correlation of 0.888 and complete-tree recovery of 87.3%, versus 0.627 and 19.0% for D_4. Held-out simulations mapped the (δμ)^2 distance to p=2 and most allele-sharing and variance-component comparators to p=1. Lower orders were more stable under finite panels and tail-producing observation errors. Order changed nearest-neighbor decisions for 5/8 Acrossocheilus fasciatus populations, 4/6 Gunnison sage-grouse populations, and 8/15 Colorado potato beetle sites. Maximum agreement with a nuclear single-nucleotide polymorphism (SNP) matrix, a double-digest restriction-site-associated DNA (ddRAD) SNP matrix, and a landscape comparator occurred at intermediate, low, and high orders, respectively; modest agreement and frequent bootstrap ties yielded no universal best order. Increasing p concentrated empirical transport on fewer loci. In targeted families, eight markers excluded from an optimized parentage panel had observed-to-Mendelian-null moment ratios of 5.17 at p=1 and 5.90 at p=4, compared with 1.29 and 1.24 for retained markers. The resulting profile provides an interpretable SSR distance report: D_1 is a stable whole-distribution anchor, and increasing p reveals decision changes and locus concentration driven by larger repeat displacements.
Items connected by shared entities, co-authorship, citations, or semantic similarity.