Runs of homozygosity (ROH) are increasingly being analyzed using whole genome sequences in non-model species as a measure of inbreeding and to assess demographic history, thus providing useful information for conservation. However, most studies have used Plink for ROH inference which has been shown to perform poorly when sequencing depth is below 10X, often underestimating the true proportion of the genome in ROH, which could lead to erroneous status assessment and management decisions. We use whole genome sequences from caribou, a non-model species at risk, subsampled to sequencing depths ranging from 1X to 15X, to assess the performance of ROHan, a program developed to enable ROH estimation using lower coverage sequences but so far only optimized for human data. We use 22 individuals with varying extent of inbreeding to assess the effects of sequencing depth, input parameters, and demographic history on the inference of ROH. We found that accurate estimation of the percentage of the genome and lengths of ROH can be achieved down to depths as low as 3-5X. However, input parameters and the demographic history of the individual can have a dramatic effect on results. Using our optimized settings, we then re-analyze low coverage sequences from a small and isolated caribou population and demonstrate high levels of inbreeding which had previously been missed. We provide recommendations for thorough optimization of parameters including the need for multiple runs as well as careful interpretation of outputs to enable robust ROH inference using low coverage whole genome sequences in wildlife species.
