Software

geno2ped: A fast and accurate R package for reconstructing pedigrees directly from SNP genotypes

---

Abstract

Accurate pedigree information remains essential for animal breeding programs but often suffers from recording errors and incomplete parentage data. With the expansion of genomic testing in cattle and other species, there is an increasing need for automated and reproducible pedigree reconstruction directly from genotype data. We developed geno2ped, an open-source R package that infers and verifies pedigrees from SNP genotypes using a deterministic kinship similarity framework that integrates opposing-homozygote counts, marker-level sharing, and vectorized Mendelian error validation. Unlike Sequoia, which depends on likelihood-based parentage assignment and heuristic genotype probabilities, geno2ped directly calculates kinship similarity scores from allele dosages and employs rule-based trio validation, making the algorithm transparent, interpretable, and readily adaptable to non-model species. The package reads standard PLINK/VCF or genotype matrix inputs and outputs a pedigree-ready file, along with summary statistics and QC plots for kinship and Mendelian error distributions. In simulated dairy-like populations (15 sires, 30 dams, 120 offspring; 1,000 SNPs), geno2ped reconstructed complete trios with 100% accuracy for both sire and dam assignment, while maintaining a Mendelian error rate below 0.005. Compared to Sequoia, geno2ped achieved identical accuracy but with 2.5 times faster runtime (approximately 1.9 seconds versus 4.9 seconds elapsed on 165 individuals, R 4.0.1, single core). Runtime scaled linearly with sample size and SNP density, and the "high-precision" preset minimized false assignments through stricter similarity thresholds and trio consistency checks. The "high-precision" preset applies stricter thresholds (e.g., similarity ≥ 0.85, ≤ 10 opposing homozygotes, and Mendelian-error ≤ 0.002), thereby prioritizing certainty over completeness by excluding ambiguous parent-offspring matches that might arise from genotyping noise or distant relatedness. Overall, geno2ped provides a transparent, rule-based alternative to probabilistic parentage inference, enabling quick, reproducible, and easily auditable pedigree reconstruction across livestock and aquaculture breeding programs.

Keywords: 2026

How to Cite:

Ampofo, I., Miglior, F., Malchiodi, F. & Fragomeni, B., (2026) “geno2ped: A fast and accurate R package for reconstructing pedigrees directly from SNP genotypes”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2284362. doi: https://doi.org/10.31274/wcgalp.23565

Rights: 1