Evaluating imputation accuracy for ancient low-coverage domestic cat genomes

Recent studies on animal domesticates (e.g., dogs, cattle, donkeys) have shown that imputing ancient low-coverage genomes can achieve high accuracy, enabling finer-scale population genomic analyses based on haplotypes. However, these studies underscore the lack of a standard imputation strategy, as species-specific factors—such as genetic architecture, introgression from wild relatives, and reference panel composition—critically influence accuracy. Despite being one of humanity’s closest companions, the domestic cat (Felis catus) remains underrepresented in genomic research, leaving many aspects of its evolutionary history unresolved. Ancient and modern genomic datasets for both wild and domestic cats are sparse, with population allele frequencies often inferred from single or few individuals, resulting in potential biases. Addressing these biases requires haplotype-based approaches and thus a tailored imputation pipeline. Imputing cat genomes poses several challenges, including the absence of a high-density genetic map essential for phasing the reference panel. Additionally, only three high-coverage genomes of Felis lybica lybica, the domestic cat’s wild ancestor, are currently available. Including more genomes of wild relatives as well as ancient high-coverage genomes in reference panels has proven to enhance imputation accuracy. To enrich the reference panel, we thus generated novel modern and ancient high-quality genomes (>10X) of both wild and domestic cats. By constructing a fine-scale genetic map and testing various imputation filtering pipelines, we aim to establish a gold standard for cat imputation, enabling robust haplotype-based analyses. This will provide unprecedented insights into the domestication, adaptation, and evolutionary history of domestic cats.

Abstract of an Oral Presentation or a Poster