The effect of the accuracy of deleterious variant prediction on genetic rescue of threatened populations

Background

Genetic rescue is a strategy for conservation of a population by introducing new genetic variation, used for example in conservation programs for populations of wild animals or in systematic crossbreeding programs for effectively small dog breeds. Genetic rescue can improve population fitness, but can also introduce new deleterious variation to an already vulnerable population.

Now that genome sequencing has become affordable and accessible for non-model organisms, researchers have proposed to use sequencing to improve genetic conservation, for example for prioritise populations or identify suitable animals for genetic rescue. These applications require that we are able to predict the function of variants from sequence data, which is a difficult.

Aims

This project aims to evaluate the impact of the accuracy of deleterious variant prediction on measurements of genetic load and genetic rescue strategies. To achieve this, the project will use simulations of deleterious variants in effectively small animal populations.

Project description

The project will use the genetic simulation environment SLiM to simulate effectively small populations with deleterious mutations, using parameters based on the literature and real data.  Predictors of variant function with different accuracies (as measured by correlation, area under the ROC curve, and enrichment for rare variants) will be simulated based on the true selection coefficients.

These predictors will be used to predict genetic load in case studies such as: simulated rescue experiments with outbred and inbred rescuers, genomic detection of purging and inbreeding depression, or simulated selection experiments.

Specifications

The project is suitable for students in animal science or bioinformatics who are interested in modelling, conservation or evolution.