Fig. 2

GWAS identification of candidate genes for variation in heterozygosity associated with key agronomic traits. a–f GWAS identification of Sc10g012040 and Sc10g012050 as candidate genes for SR epidermal type on chromosome 10. a Manhattan plots for SR epidermal type on chromosome 10 using FAST-LMM. Red arrow indicates the significant GWAS peak. b Local manhattan plot (top) and linkage disequilibrium heat map (bottom) surrounding the GWAS signal. Red dashed lines indicate the candidate region. Red dots indicate the core SNPs in the candidate gene. c,d Sc10g012040 and Sc10g012050 gene model, respectively. The dashed red lines represent the position of nonsynonymous SNPs. e Images of rough and smooth types of SR epidermis. f Comparison of SR epidermal types based on the haplotypes in Sc10g012040 (left) and Sc10g012050 (middle), as well as in both Sc10g012040 and Sc10g012050 (right). Hap 1 and hap 3 indicate the homozygous allele combinations, while hap 2 shows the heterozygous allele combinations. The symbol n represents the number of accessions with the same genotype. g–i GWAS identification of Sc03g001750 as a candidate gene for SR epidermal type on chromosome 3. j–m GWAS identification of Sc13g000920 as a candidate gene for SR scar on chromosome 13. n–p GWAS identification of Sc05g013530 as a candidate gene for above-ground weight of cassava plants on chromosome 5. q–s GWAS identification of Sc11g000910 as a candidate gene for SR amylopectin content on chromosome 11. t–v GWAS identification of Sc18g013220 as a candidate gene for cassava resistance to T. cinnabarinus on chromosome 18. For box plots in (p), (s), and (v), the center line represents the median, box limits indicate the upper and lower quartiles, and whiskers denote the range of the data. The significance of difference was derived with two-tailed t-test (*P< 0.05, **P< 0.01). For histograms in (f), (i), and (m), the significance of difference was derived with the chi-square test (*P< 0.05, **P< 0.01)