Identification of Favorable Alleles from Exotic Upland Cotton lines for Fiber Quality Improvement using Multiple Association Models

Hrithik Mangla ¹ Min Liu ¹ Deepak Vitrakoti ¹ Rama Vamsi Somala ¹ Tariq Shehzad ² Rahul Chandnani ¹ Sayan Das ³ Jason G Wallace ¹ John L Snider ¹ Don Jones ⁴ Peng Chee ¹ Andrew H Paterson ^1*

• ¹ University of Georgia, Athens, United States
• ² Symmetry biosciences, Durham, United States
• ³ Corteva Agriscience (India), Hyderabad, Andhra Pradesh, India
• ⁴ Cotton incorporated, Cary, North Carolina, United States

Upland cotton (Gossypium hirsutum) faces the challenge of limited genetic diversity in the elite or improved gene pool. To address this issue, we explored alleles contributed by five 'converted' exotic lines sampling most of the undomesticated botanical races of G. hirsutum, in BC1F2 and F3 populations. Joint analysis of all populations along with population-specific analyses identified 38 unique QTL for six different fiber quality traits. At 15 of these loci, DES56 or the elite allele improved upon all the exotics. For another 15, only a single of the five exotics improved upon the elite allele, suggesting the rare alleles that may not have been sampled in the cotton domestication or improvement. At the remaining 8 QTL, multiple exotic lines contributed the superior allele, suggesting that DES56 (and by extension the elite gene pool) has chronically poor alleles at these loci. Converted strains T1046, T326, and T063 showed the highest potential for contributions to cotton fiber quality breeding programs. Upper Half Mean Length and Fiber Strength showed multiple QTL regions affecting both traits simultaneously, while the Uniformity Index showed the smallest heritability values. The estimation of pairwise genetic distances for six parental lines indicates that DES56 has a higher genetic similarity with each exotic line than the exotic lines have with each other. Most of the detected QTL were 'minor' (explaining less than 10% of variance) supporting the implementation of genomic selection techniques to utilize the cumulative effects of most of these QTL distributed genome-wide. Finally, some regions were consistently unfavorable for exotic introgression such as on chromosomes A13 and D09, indicating the possible genomewide haplotypes that may combine the benefits of a history of scientific breeding of the elite gene pool.