Genome-wide association analysis of anthracnose resistance in sorghum [Sorghum bicolor (L.) Moench]

Abstract

In warm-humid ago-ecologies of the world, sorghum [Sorghum bicolor (L.) Moench] production is severely affected by anthracnose disease caused by Colletotrichum sublineolum Henn. New sources of anthracnose resistance should be identified to introgress novel genes into susceptible varieties in resistance breeding programs. The objective of this study was to determine genome-wide association of Diversity Arrays Technology Sequencing (DArTseq) based single nucleotide polymorphisms (SNP) markers and anthracnose resistance genes in diverse sorghum populations for resistance breeding. Three hundred sixty-six sorghum populations were assessed for anthracnose resistance in three seasons in western Ethiopia using artificial inoculation. Data on anthracnose severity and the relative area under the disease progress curve were computed. Furthermore, the test populations were genotyped using SNP markers with DArTseq protocol. Population structure analysis and genome-wide association mapping were undertaken based on 11,643 SNPs with <10% missing data. The evaluated population was grouped into eight distinct genetic clusters. A total of eight significant (P < 0.001) marker-trait associations (MTAs) were detected, explaining 4.86-15.9% of the phenotypic variation for anthracnose resistance. Out of which the four markers were above the cutoff point. The significant MTAs in the assessed sorghum population are useful for marker-assisted selection (MAS) in anthracnose resistance breeding programs and for gene and quantitative trait loci (QTL) mapping.

Fig 1. Rainfall, relative humidity, and minimum and maximum temperatures of Bako Agricultural Research Center in 2016 (A), 2017 (B) and 2018 (C).

(Source: Bako Agricultural Research Center/Ethiopia).

Fig 2. The relative area under disease progress curves (rAUDPC) of 10 anthracnose resistant and 10 susceptible sorghum genotypes, including Btx623 among 366 genotypes evaluated in three seasons in Ethiopia.

Fig 3. Plot depicting the cross-validation error rates values and K sub-sets varying from K = 2 to K = 20 based on ADMIXTURE analysis.

Fig 4. Population genetic structure among 313 Ethiopian sorghum accessions in K = 8.

Fig 5. Principal component analysis among 313 sorghum collections based on 11,643 SNPs using the first two principal components.

The large proportions of the variances contained in the data are retained by the first three principal components (A), while the relationship among collections between areas of origin is represented in (B).

Fig 6. Genome-wide association of anthracnose resistance amongst 313 sorghum collections with 11,643 SNP markers using a FarmCPU Model: (a) Manhattan plots showing significant false discovery rate (FDR) adjusted P-value of ≤0.1 associated with anthracnose resistance, A dash line represents the threshold from the FDR, and a blue line represents the significant threshold −Log10 (P) value and (b) Log Q-Q plots validating the FarmCPU Model and depicting consistency in reducing -log10(p-values) toward the expected level.