Landraces (traditional varieties) of domesticated species preserve useful genetic variation, yet they remain untapped due to the genetic linkage between the few useful alleles and hundreds of undesirable alleles. We integrated two approaches to characterize the diversity of 4,471 maize landraces. First, we mapped genomic regions controlling latitudinal and altitudinal adaptation and identified 1,498 genes. Second, we used F-one association mapping (FOAM) to map the genes that control flowering time, across 22 environments, and identified 1,005 genes. In total, we found that 61.4% of the single-nucleotide polymorphisms (SNPs) associated with altitude were also associated with flowering time. More than half of the SNPs associated with altitude were within large structural variants (inversions, centromeres and pericentromeric regions). The combined mapping results indicate that although floral regulatory network genes contribute substantially to field variation, over 90% of the contributing genes probably have indirect effects. Our dual strategy can be used to harness the landrace diversity of plants and animals.

In the version of this article initially published online, there were two errors. In the section "Three classes of monoallelic elements" in the main text, "We classified all monoallelically accessible elements (1,966 elements)" should have read "1,964 elements. " In the legend for Figure 5c, the number of elements open in ESCs should have been given as 234 instead of 35.

The objective of this study was to evaluate Maize (Zea mays L.) elite lines currently available in CIMMYT's lowland tropical breeding program in Latin America under multiple abiotic stresses and identify lines with tolerance to drought, N deficiency, and combined heat and drought stress (HTDS). An incomplete line‐by‐tester design was used to evaluate 436 testcrosses under nonstressed conditions, 507 under N deficiency, 417 under drought stress (DS), and 368 under HTDS in 30 season‐by‐location combinations between 2012 and 2015. Elite lines CLRCY016, CML269, CML550, and CML551 performed well across all conditions, while CLQRCWQ118, CLWN306, and CML576 showed good performance under DS and N deficiency. CML574 was tolerant to DS and HTDS. Moreover, CML550 and CML574 are known for their partial tolerance to maize lethal necrosis. Grain yield measured under DS was to some extent predictive of attainable grain yield under N‐deficient conditions (r = 0.65; P < 0.01) and HTDS (r = 0.54; P < 0.01) as indicated by the correlation across treatments. The fact that only a few lines were tolerant across treatments re‐emphasizes the need to separately screen germplasm under each abiotic stress. Based on high best linear unbiased predicted general combining ability (BLUP GCA), it will be possible to develop hybrids tolerant to multiple abiotic stresses without incurring any yield penalty under nonstressed conditions using these inbred lines.

No abstract