Genetic analysis of resistance to african cassava whitefly, bemisia tabaci in crosses of latin american by Ugandan cassava parental lines

Abstract

The African Cassava whitefly, Bemisia tabaci is both a viral disease vector and pest to cassava, causing direct damage through adults and nymphs piercing and sucking behavior and whose infestations can cause up to 50% crop yield reduction. As a staple crop for more than 800 million people, most of whom are in sub-Saharan Africa, cassava (Manihot esculenta Crantz) improved with pest resistance, notably whiteflies could have enormous impact on farmers’ yields since declines in cassava yields can lead to household food, nutrition, and income insecurity. As a result of this, disease and whitefly tolerant/resistant genotypes from Latin America and regional elites were introduced to Uganda to help improve on selected local germplasm. The objectives of this study were to (1) assess population structures among crosses of introduced Latin American cassava germplasm with local adapted progenitors, (2) identify genomic regions in cassava associated with resistance to African cassava whitefly, and (3) assess the relationship between the cassava whitefly resistance and other important cassava agronomic traits amongst the selected cassava varieties. Data analysis was done using single nucleotide polymorphic genotyped data together with phenotypic data from a panel of 750 clones. For objective 2, a genome-wide association study (GWAS) was conducted to uncover genomic regions associated with natural resistance to African Cassava whitefly, while objective 3 assessed phenotypic correlations between whitefly resistance traits, and other agronomic traits (disease (CMD and CBSD) intensity, yield and DMC). A total of 15,625 single nucleotide polymorphism markers were used to assess genetic diversity. The SNP markers were informative, with polymorphic information content ranging from 0.35 to 0.41 and a mean of 0.38 across all the genotypes. The observed heterozygosity ranged from 0.26 to 0.36 with a mean of 0.32 whilst the expected heterozygosity ranged from 0.29 to 0.36 and the mean was 0.32, and the minor allele frequency ranged from 0.19 to 0.26 with an average of 0.23. The principal coordinate analysis, model-based structure and the Euclidean distance matrix method grouped the 736 genotypes into four main clusters irrespective of origin. This indicates that the genotypes analyzed in this study may have shared common alleles across the population. The genotypes employed in this study also exhibited a moderate level of genetic variation. Analysis of molecular variance revealed that within-population variation among the seven genetic groups based on pedigree accounted for 92% of the observed variation. The locally adapted progenitors from Uganda were found to have greater genetic diversity in comparison to the imported progenitors from Latin America. A total of 9 significant single nucleotide polymorphisms (SNPs) markers were identified to be associated with the nymph count trait on nine chromosomes; 2, 4, 8, 10, 11, 12, 13, 15 and 17. Gene annotation analysis revealed the presence of thirty five genes subdivided into categories according to their protein structure namely; pentatricopeptide, protein kinase, zinc finger, 2-oxoglutarate (2OG) and Fe (II)-dependent oxygenase, zinc knuckle family proteins, and auxin-responsive transcription factor genes, which collectively are involved in regulation of a wide variety of functions ranging from developmental and plant defense-related responses to environmental as well as pathogen attack. The results obtained from this work revealed significant SNP variants and genes that may be utilized for genomic-informed selection in order to enhance whitefly resistance in cassava. We observed the highest phenotypic correlation (r = 0.82) between adult whitefly count and nymph count, suggesting the usefulness of single trait data in predicting performance of related traits. The observed high phenotypic correlations between adult whitefly count and nymph count indicate that data obtained on either trait would suffice.