Genome-wide association analysis for heat tolerance at flowering detected a large set of genes involved in adaptation to thermal and other stresses

Abstract

Fertilization sensitivity to heat in rice is a major issue within climate change scenarios in the tropics. A panel of 167 indica landraces and improved varieties was phenotyped for spikelet sterility (SPKST) under 38°C during anthesis and for several secondary traits potentially affecting panicle micro-climate and thus the fertilization process. The panel was genotyped with an average density of one marker per 29 kb using genotyping by sequencing. Genome-wide association analyses (GWAS) were conducted using three methods based on single marker regression, haplotype regression and simultaneous fitting of all markers, respectively. Fourteen loci significantly associated with SPKST under at least two GWAS methods were detected. A large number of associations was also detected for the secondary traits. Analysis of co-localization of SPKST associated loci with QTLs detected in progenies of bi-parental crosses reported in the literature allowed to narrow -down the position of eight of those QTLs, including the most documented one, qHTSF4.1. Gene families underlying loci associated with SPKST corresponded to functions ranging from sensing abiotic stresses and regulating plant response, such as wall-associated kinases and heat shock proteins, to cell division and gametophyte development. Analysis of diversity at the vicinity of loci associated with SPKST within the rice three thousand genomes, revealed widespread distribution of the favourable alleles across O. sativa genetic groups. However, few accessions assembled the favourable alleles at all loci. Effective donors included the heat tolerant variety N22 and some Indian and Taiwanese varieties. These results provide a basis for breeding for heat tolerance during anthesis and for functional validation of major loci governing this trait.

Fig 1. Unweighted neighbour-joining tree based on simple matching distances constructed from genotype of 201 accessions of the indica diversity panel, using 825 SNP markers.

Subpopulation 1: traditional lowland indica; Subpopulation 2: improved lowland indica; Subpopulation 3: traditional lowland varieties from Madagascar; Subpopulation 4: aus accessions; m: admixed accessions. Subpopulation numbers followed by “P” refer to accessions that were phenotyped in the framework of the present study.

Fig 2. Projection of the contribution of the 20 phenotypic traits on the plan of the first two axes of a Principal Component Analysis.

Traits subjected to three methods of GWAS are indicated in blue.

Fig 3. Distribution of spikelet sterility rate among the four Structure subpopulations composing the diversity panel.

SP1: traditional lowland indica; SP2: improved lowland indica; SP3: traditional lowland varieties from Madagascar; SP4: aus accessions; m: admixed.

Fig 4. Co-localization of significant SNPs and haplotypes detected by the three GWAS methods with QTL for heat tolerance during the reproductive phase detected in biparental crosses.

Blue: Sm-GWAS, Green: Hap-GWAS, Black: BL-GWAS and Red: QTL.