Flowering time regulation model revisited by pooled sequencing of mass selection populations

Maize is one of the most broadly cultivated crops throughout the world, and flowering time is a major adaptive trait for its diffusion. The biggest challenge in understanding maize flowering genetic architecture is that the trait is confounded with population structure. To eliminate the effect, we revisited the flower time genetic network by using a tropical maize population Pop32, which was under mass selection for adaptation to early flowering time in China for six generations from tropical to temperate regions. The days to anthesis (DTA) of the initial (Pop32C0), intermedia (Pop32C3), and final population (Pop32C5) was 90.77, 84.63, and 79.72 days on average, respectively. To examine the genetic mechanism and identify the genetic loci underlying this rapid change in flowering time of Pop32, we bulked 30 individuals from C0, C3, and C5 to conduct the whole genome sequencing. And we finally identified 4,973,810 high-quality single nucleotide polymorphisms (SNPs) and 6,517 genes with allele frequency significantly changed during the artificial improvement process. We speculate that these genes might participate in the adaptive improvement process and control flowering time. To identify the candidate genes for flowering time from the gene set with allele frequency changed, we carried out weighted gene co-expression network analysis (WGCNA), and identified four co-expression modules that highly associated with the flowering time development, as well as constructed the co-expression network of key flowering time genes. Gene Ontology (GO) enrichment analysis revealed that the GO terms photosynthesis/light reaction, carbohydrate binding, auxin mediated signaling pathway, response to temperature stimulus that are closely connected with flowering time. Furthermore, targeted GWAS revealed the genes are significantly connected with the flowering time. qRT-PCR of four candidate genes GRMZM2G019879, GRMZM2G055905, GRMZM2G058158, and GRMZM2G171365 showed that their expression level is similar to the flowering time genes, which playing a key role in maize flowering time transition. This study revealed that the changes of flowering time in mass selection process may be strongly associated with the variations of allele frequency changes, and we identified some important candidate genes for flowering time, which will provide a new insight for the rapid improvement of maize important agronomic traits and promote the gene cloning of maize flowering time.