Cassava (Manihot esculenta Crantz) is a starchy root crop cultivated by up to 800 million people in the tropics (Nasser and Ortiz, 2010). The crop has multiple uses for food, feed and industry. There is use for almost every part of the cassava plant, but the most economically important part are its starchy roots.  However, the usefulness of cassava roots is limited by presence of the endogenous cyanogenic glucosides (CGs). Linamarine and lotaustrine are the commonest CGs in cassava but linamarine is the most abundant (Gleadow & Møller, 2014). The linamarine is readily hydrolyzed to glucose and acetone cyanohydrin in the presence of linamerase enzyme, which is also inherently produced by the plant. Acetone cyanohydrin decomposes rapidly in neutral or alkaline solution, liberating cyanide (HCN) gas.

Cassava with HCN content > 100ppm is said to be bitter while that with more than 10 ppm on dry weight basis is unsafe for human consumption (FAO/WHO, 2001 ). High HCN not only makes cassava bitter but is also toxic to lethal levels (Alitubeera et al., 2019) while chronic exposure to even low concentrations of dietary HCN is linked to several debilitating health conditions including the permanent, irreversible paralytic conditions konzo and tropic ataxic neuropathy (TAN) (Teles, 2002;Nhassico et al., 2008;Cliff, 2011;Cliff et al., 2011). Other conditions include goiter and cretinism. Whereas some communities in Uganda deliberately cultivate bitter cassava varieties citing security from mammalian pests including humans, rodents and livestock, the vast majority of end-users prefer non-bitter cassava varieties characterized by low HCN content. 

Progress in breeding for low HCN cassava accessions has been slow, mainly owing to the difficulty in HCN quantification as available laboratory methods like the enzymatic method, acid hydrolysis and picrate are slow, laborious, time consuming. The methods are limited in throughput and/or accuracy and therefore cannot be reliably used to screen large early stage gate seedling and clonal trials where diversity for the trait is maximum. Thus, it is difficult to identify parents to send to the crossing blocks for low HCN breeding. Faster and more reliable methods of HCN screening (phenotyping) are desirable.

The cassava breeding community has embraced  of NIR infrared spectroscopy (NIRS) as a fast, accurate and reproducible tool for quantification of cassava root constituents (Alamu et al., 2021; Abincha et al., 2022). Thus NIRS has successfully been deploy in determination of root dry matter content, starch content, carotenoid content and amylose content. On the other hand, cassava breeders have moved fast to adopt genomic tools in cassava breeding. Thus, modern cassava breeding programs are deploying genomic selection (GS) and Marker Assisted Selection (MAS) in routine breeding work to shorten breeding cycles, increase accuracy of selections and thus ensure faster genetic gains. The concept of GS has been proven for cassava brown streak disease (CBSD) (Kayondo et al., 2019; Ozimat et al., 2019; Ano et al., 2021) and HCN (Torres et al., 2021) while MAS has been applied for cassava mosaic disease (CMD) (Ige et al., 2021)

My study seeks to; 1)determine the extent of genetic variation for HCN in Ugandan cassava germplasm, 2) validate Kompetitive Allele Specific PCR (KASP) markers in a Ugandan population, 3) evaluate the prediction accuracy of NIRS for fresh cassava root HCN content, 4) evaluate the accuracy of genomic prediction models for HCN.

For any selection to be made and indeed for NIRS and GS, there must be genetic variation within the study population. Validation of KASP markers will be the first step in marker assisted selection for HCN, NIRS will provide a fast, accurate and reliable method for HCN phenotyping. Overall, it is hoped that this study will enhance breeding for end-user preferred cassava varieties by developing faster, accurate and reproducible HCN quantification methods to inform selection decisions by the national cassava breeding program.