Inference of Maize Population History during Migration to Highland Habitats

An understanding of maize population history is significant for the study of maize evolution. However, little is known about the demographic history of maize following its initial domestication. Colonization of high altitude environments occurred multiple times during the diffusion of maize populations. A founder effect (bottleneck) is hypothesized to have occurred in highland populations during colonization and adaptation to novel environmental conditions, associated with divergence from their lowland progenitors. However, the timing of colonization and bottlenecks as well as the intensity of bottlenecks remain unknown. The emergence of new population history inference methods, which explicitly model ancestral relationships under recombination and mutation, enables us to infer ancestral effective population size and even population separations. Here, we sampled four individuals from four highland populations (the southwestern United States, the Guatemalan highlands, the Mexican Central Plateau and the Andes of South America) and two lowland populations (from Mexico and South America). Genomes of the 24 samples were deeply sequenced (average depth varying from 20X to 40X). Raw reads were aligned to the maize B73 reference genome (RefGen_v3) with BWA, and then realigned bam files were cleaned with Picardtools (Mark Duplicates) and GATK (Indel Realignment). Subsequently, three methods, PSMC (Pairwise Sequentially Markovian Coalescent), diCal (Demographic Inference using Composite Approximate Likelihood) and MSMC (Multiple Sequentially Markovian Coalescent), were utilized to infer effective population size change over history. These analyses provide insights into the demographic history of maize populations during early diffusions into highland environments.