Mapping the Genomic Limits of De-Extinction in the Face of Ancient DNA Degradation

The de-extinction of species using genome-editing approaches depends on acquiring high-quality genomic information from the extinct target. However, the degraded nature of the ancient DNA (aDNA) that is typical for most extinct species, poses significant challenges to achieving comprehensive genome reconstruction. A systematic evaluation of the minimum sequencing effort that is required to reliably map the genome under varying DNA quality conditions to different reference genome remains lacking across different extinct species. Here, we systematically assess the impact of sequencing depth on genome coverage, heterozygosity estimation, and variant calling accuracy, when mapping both true aDNA data generated from the extinct Christmas Island rat (Rattus macleari), as well as in silico simulated modern- and ancient-like data generated from a modern relation (the brown rat, Rattus norvegicus), to the black rat (Rattus rattus) reference genomes. Our results demonstrate that even sequencing depths of 100× fail to yield stable heterozygosity estimates, and leave approximately 3.38% to 4.03% of its genome uncovered. These uncovered regions contained functionally relevant SNPs and indels, highlighting the limitations of reconstructing extinct genomes using reference sequences from extant relatives. Furthermore, simulations using computationally generated "degraded haploid and diploid" data based on the high-quality brown rat genome, revealed that false-positive SNPs primarily arise from insufficient coverage and low data quality, rather than aDNA damage (e.g. miscoding lesions, size of fragments, etc.) per se. These findings underscore the need to tailor sequencing depth standards by considering sample type, degradation level, and sequencing error profiles. This study provides a theoretical framework and methodological support for optimizing data strategies in aDNA research, and ultimately informing de-extinction efforts.