Using approximate Bayesian computation to infer ancestral population size of pre-Columbian North American human populations
Advances in computational algorithms provide an opportunity to examine the demographic history of populations, i.e., whether they have grown or shrank over time. We would like to investigate the history of humans in North America, before European contact to see how many people lived there and how the population changed over time. While a comprehensive data set exists, it is genotyped on an array, which creates biases that can throw off the existing computational methods, which work with more modern sequencing data. We will modify an existing software package (PopSizeABC) to account for these biases. We will test the software using data on cows, where we have good data from next-generation sequencing, and extensively studied population history. This project requires familiarity with R and Python, or a good computational background that will allow you to pick them up quickly.