Item

Abstract

Four thousand years ago, maize began migrating from Mexico to the geographically diverse Southwestern US. Over 2,000 years of population selection, early Puebloan people adapted this germplasm to the short- season uplands. We show that the resulting germplasm became the Northern Flints, which provide temperate adaptation to modern US, Chinese, and European varieties. We combined mapping and genomic prediction trained on large modern inbred populations with population analyses from modern landraces and ancient DNA from the Southwest to understand the biological basis, timing, and population dynamics underlying early flowering. We addressed the biological basis of early flowering using six modern populations, totaling nearly 15,000 lines. We remapped flowering on 81 million projected whole-genome sites using published SNP-based and novel variance component approaches, and combined results across populations in a meta-analysis. We identified significant regions resulting from the Southwest adaptation by comparing results to Fst calculated between Northern Flint and tropical lines. Enrichment in functional, evolutionary conservation and population genetic annotations underlying significant regions elucidates the biology underlying flowering time, clarifying the mechanisms of adaptation. We investigated the population dynamics underlying early flowering with GBS-genotyped and partially phenotyped modern SW-focused landraces and 14 whole-genome sequenced cobs from the Turkey Pen archaeological site, an early intensifying-agricultural site in the uplands. Population analyses of Turkey Pen and modern landraces suggest that Turkey Pen is ancestral to both the Northern Flints and modern Southwestern landraces. We validated the flowering model trained on modern inbreds with our phenotyped landrace population, demonstrating good prediction accuracy, and applied it to Turkey Pen to estimate extent of adaptation. We also present segregation results for known early flowering loci in Turkey Pen, such as the MITE insertion at vgt1. These data suggest that early flowering was mostly in place by 2,000 years ago, and relied on standing variation from lowland north Mexico.