Sequence-aware genotyping improves the accuracy of molecular markers associated with cassava mosaic disease resistance loci

Cassava ( Manihot esculenta ) is a critical food crop for sub-Saharan Africa. Cassava mosaic disease (CMD), caused by cassava geminiviruses, can severely reduce crop yield. Historically, CMD-resistance breeding has played a pivotal role in CMD management. Marker-assisted selection (MAS) for three CMD-resistance–associated quantitative-trait-loci (QTLs) (CMD1, 2 and 3) formed the basis for many elite cassava breeding lines currently deployed across Africa. Six molecular markers linked to these loci (SSRY40, SSRY28, NS158, NS169, RME1, and NS198) are still employed in contemporary CMD resistance breeding. False discovery in MAS for CMD resistance is frequently attributed to epistasis and recombination, but the contribution of size homoplasy remains under-researched. Sequencing and mapping of CMD-linked marker loci (MLs) revealed a substantial (10–20%) and significant ( p < 0.05) rate of homoplasy-induced false discovery. Furthermore, mapping revealed CMD-linked MLs in genic regions. Four putative CMD-response genes ( PDCD7 , HMGB1 , PP2A-B’δ and LOC128695772) were identified through transcriptional and ontological analyses of genes containing CMD-linked MLs. Based on a small but phylogenetically diverse collection of cassava germplasm, this study demonstrates the potential of sequence-aware marker genotyping to improve the efficiency of CMD-resistance breeding. Marker non-neutrality was identified as an additional source of error in CMD QTL-mapping and screening.