The missing data problem in population genomics and statistical methods to address them

The "Missing Data" problem is prevalent across all statistical inference, owing to the "absence of some part of a familiar data structure" (Efron 1994).Population genomic datasets are riddled with missing data (Fig. 1)-broadly classified as data missing at random (e.g.due to degradation, sequencing errors), data missing "on purpose" (e.g.due to sequencing strategies like genotyping by sequencing), and data missing due to unknown evolutionary history (e.g.introgression from ancestral ghost populations).Editors and scientific contributors to both the GSA's journals, Genetics and G3 have continually highlighted statistical issues and pitfalls with inference in the presence of missing data (McIntyre 2025), particularly in an age of Biobank scale population genomic datasets.Here I highlight studies, including those that have been recently published in Genetics and G3 towards systematically assessing the effects of missing data problems and addressing them towards inference in a variety of population genomics questions.