Genome‐Wide Association and Prediction Reveals Genetic Architecture of Cassava Mosaic Disease Resistance and Prospects for Rapid Genetic Improvement

C ( Crantz) is a crucial, under-researched crop feeding millions worldwide, especially in Africa. Cassava mosaic disease (CMD) has plagued production in Africa for over a century. Biparental mapping studies suggest primarily a single major gene mediates resistance. To investigate this genetic architecture, we conducted the first genome-wide association mapping study in cassava with up to 6128 genotyping-by-sequenced African breeding lines and 42,113 reference genome-mapped single-nucleotide polymorphism (SNP) markers. We found a single region on chromosome 8 that accounts for 30 to 66% of genetic resistance in the African cassava germplasm. Thirteen additional regions with small effects were also identified. Further dissection of the major quantitative trait locus (QTL) on chromosome 8 revealed the presence of two possibly epistatic loci and/or multiple resistance alleles, which may account for the difference between moderate and strong disease resistances in the germplasm. Search of potential candidate genes in the major QTL region identified two peroxidases and one thioredoxin. Finally, we found genomic prediction accuracy of 0.53 to 0.58 suggesting that genomic selection (GS) will be effective both for improving resistance in breeding populations and identifying highly resistant clones as varieties.