Sequencing & Pangenomes

Effects of sequencing depth on the reliability of heterozygosity estimation in pigs

---

Abstract

IntroductionGerman native pig breeds, such as the Swabian-Hall, are increasingly threatened with extinction due to the shift to highly intense production schemes. Despite this, these breeds are expected to offer significant advantages, particularly increased resilience to changing environments, such as extreme weather or non-conventional feed. This makes them particularly valuable in organic farming, where animals are often exposed to free-range systems and alternative feeds.Genomic characterization is generally accepted to be beneficial in conservation breeding schemes and the current standard are genotyping arrays. Low-coverage Long-read sequencing offers a flexible and fast alternative at reasonable costs, if reliable diversity estimates can be achieved with low depth of coverage (DoC).ObjectivesThis study aims to investigate whether lower sequencing depths can still provide reliable estimates for relevant parameters, such as heterozygosity, and to determine if these estimates remain stable and accurate across different sequencing depths.Material and methodsWe used Oxford Nanopore long-read sequencing to sequence eight Swabian-Hall pigs to DoC of about 10 to 15 X. Raw data was base-called and mapped with a Dorado/ Minimap2 based pipeline. The animals were subsequently down-sampled to DoC´s of 10, 5, 2, 1, 0.5 and 0.2 X, respectively. ANGSD was employed to estimate site allele frequencies and derive the site frequency spectrum, independently for each individual. As an exemplary measure of diversity, we estimated individual heterozygosity as the expected proportion of heterozygous sites, calculated by normalizing the k = 1 class of the site frequency spectrum inferred with ANGSD/realSFS, for every sample and down-sampling step.ResultsHeterozygosity at 10 X ranged from 0.12 % to 0.22 % over whole chromosome 1. Down-sampling led to a nearly linear decrease in estimated observed heterozygosity of 20 to 40 % at DoC of 1 X. At lower DoC, observed heterozygosity estimates were unstable and even showed increased values compared to higher DoC. The rate of decrease followed the same pattern for all samples. Spearman rank correlations between different DoC´s were moderate to high, around 0.95 between adjacent DoC´s and still 0.71 between 10 X and 0.5 X. Rank correlation dropped drastically between 0.5 and 0.2 X supporting the finding that heterozygosity estimates at 0.2 X appeared unstable.ConclusionWe know from a parallel study (Abstract by Reimer et. al.) that genotypes can reliably be called from low-coverage data, using a low-coverage imputation strategy, if a sufficient number of animals is sequenced. There it was also found, that genotype concordance decreased drastically below 1 X. This approach demonstrates that, even if animals are analysed completely independently, the resulting heterozygosities will allow a correct ranking of animals, helpful e.g. in a gene bank candidate selection scenario, down to a DoC as low as 1 X.

Keywords: 2026

How to Cite:

geibel, J., Hinrichs, D., Kistner, L., Petig, M., Reimer, C., Vollmer, L., Weigend, A., Weigend, S., Weishaar, R. & Worat, C., (2026) “Effects of sequencing depth on the reliability of heterozygosity estimation in pigs”, World Congress on Genetics Applied to Livestock Production Digital Archive 2026(1): 2287145. doi: https://doi.org/10.31274/wcgalp.24222

Rights: 1