Abstract

Cassava is an important crop for food security in the tropics where its production is jeopardized by several viral diseases, including the cassava mosaic disease (CMD) which is endemic in Sub-Saharan Africa and the Indian subcontinent. Resistance to CMD is linked to a single dominant locus, namely CMD2. The cassava genome contains highly repetitive regions making the accurate assembly of a reference genome challenging. In the present study, we generated BAC libraries of the CMD- susceptible cassava cultivar (cv.) 60444 and the CMD-resistant landrace TME3. We subsequently identified and sequenced BACs belonging to the CMD2 region in both cultivars using high-accuracy long-read PacBio circular consensus sequencing (ccs) reads. We then sequenced and assembled the complete genomes of cv. 60444 and TME3 using a combination of ONT ultra-long reads and optical mapping. Anchoring the assemblies on cassava genetic maps revealed discrepancies in our, as well as in previously released, CMD2 regions of the cv. 60444 and TME3 genomes. A BAC guided approach to assess cassava genome assemblies significantly improved the synteny between the assembled CMD2 regions of cv. 60444 and TME3 and the CMD2 genetic maps. We then performed repeat-unmasked gene annotation on CMD2 assemblies and identified 81 stress resistance proteins present in the CMD2 region, amongst which 31 were previously not reported in publicly available CMD2 sequences.