Animals have always been exposed to mass extinctions, but during the last few hundred years, human activities have been one of the main reasons for dramatic decline. When a population becomes smaller, threats affecting the species’ genome become larger. It becomes more prone to inbreeding, good genetic variation will be lost and damaging variation will increase. This leads to a worse ability to adapt to changing environments and then extinction. Hence, there is a crucial need to characterize and quantify damaging variation and its contribution to species decline, both for conservational purposes and understanding what drove former species to extinction. In the proposed three-year project, I will first create a bioinformatic method to score genetic variants based on their predicted deleteriousness. I will make use of sequenced genomes of pig, and annotation methods, to annotate genomic features that can be used to predict deleteriousness. In the second part, I will translate the predicted deleteriousness per variant to genetic load per individual. I will then train it within genetic data from pig, and validate it in sequence data from wild boar. Lastly, I will test the method in endangered pig species and in already extinct animals, like mammoths and rhinos. As these types of scores are almost always species specific, there is a crucial need to expand beyond species borders with the long-term goal of estimating how much genetic variation is contributing to animal extinction.