Genetic gain & Inbreeding

Comparing inbreeding parameters in pigs: Drift, homozygosity, and IBD

---

Abstract

The goal of this study was to investigate and compare different estimates of inbreeding rates in a dam line based on genomic and pedigree data, using three classes of approaches: identical by descent (IBD), drift, and homozygosity-based estimates. This study utilized a dataset comprising 34,126 purebred animals born between 2014 and 2024, consisting of nucleus sires and their genotyped relatives. The inbreeding coefficients were estimated using different approaches. IBD-based measures were obtained from the pedigree (PED) and the Genomic Relationship Matrix based on genetic linkage analysis (GLA). Drift-based measures were assessed through genomic relationship matrices (G) constructed according to VanRaden (2008; methods 1 (VR01) and 2 (VR02)) and Yang et al. (2010; YANG). Homozygosity-based measures were estimated using Runs of Homozygosity (ROH), excess of Homozygosity (HOM), and the G with the allele frequency of the markers fixed at 0.5 (GVR05). ROH and HOM were performed using a 50k and a 660k SNP panel, to compare also the relevance of SNP density. However, due to computational constraints, a 50K marker panel was used to construct the G matrix. The reference population was defined as the animals born in the year 2014. Methods were compared based on Pearson correlations and the slope of the regression from 2014 to 2024. Results showed that ROH predicted the highest average inbreeding level at 0.21 (50k) and 0.11 (660k), whereas PED and GLA had the lowest at 0.07. Drift inbreeding slopes were the highest among all methods, ranging from 0.62% per year (VR01) to 1.00% per year (VR02). The slopes of IBD inbreeding measurements were consistent at 0.38% and 0.42% per year for the PED and GLA methods, respectively. In contrast, the slopes of homozygosity inbreeding measurements exhibited a much larger variability, ranging from 0.21% per year for the ROH method to 0.42% per year for the HOM method, both measured using the 50k SNP panel. The correlations within each class of estimation methods (drift, homozygosity, or IBD) were positive and strong (ranging from 0.78 to 1.00), suggesting that the methods within a class were able to capture the same information. Additionally, the correlations between the 50K and 660K panels were strong: 0.88 for ROH and 0.95 for HOM. However, the correlations between the classes of approaches were weak (ranging from -0.47 to 0.66). Thus, it is confirmed that different classes capture different information, and that, for example, measures of homozygosity do not effectively capture drift information, and vice versa. These discrepancies between drift, homozygosity, and IBD-based measures of inbreeding also indicate that allele frequency changes were not neutral, as under a neutral model of inheritance, drift, homozygosity, and IBD-based inbreeding are expected to be equivalent.

Keywords: 2026

How to Cite:

Kelles, K., Leite, N., van Son, M., Wientjes, Y., Grindflek, E., Berg, P. & Meuwissen, T., (2026) “Comparing inbreeding parameters in pigs: Drift, homozygosity, and IBD”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2283094. doi: https://doi.org/10.31274/wcgalp.23482

Rights: 1