Jordan Li Cahoon

This paper introduces a likelihood-based framework called gLike to infer demographic history from genealogical trees, showing improved accuracy in estimating population parameters compared to existing methods.

The paper evaluates genotype imputation accuracy across global populations and finds disparities in performance for non-European ancestries, highlighting the need for more diverse reference panels.

The paper identifies inverted genomic regions between human reference genome builds that affect imputation accuracy and association testing power, proposing a heuristic to correct these issues.