Population structure and genetic diversity of the Pee Dee cotton breeding program

Abstract

Accelerated marker-assisted selection and genomic selection breeding systems require genotyping data to select the best parents for combining beneficial traits. Since 1935, the Pee Dee cotton germplasm enhancement program has developed an important genetic resource for upland cotton (Gossypium hirsutum L.), contributing alleles for improved fiber quality, agronomic performance, and genetic diversity. To date, a detailed genetic survey of the program’s eight historical breeding cycles has yet to be undertaken. The objectives of this study were to evaluate genetic diversity across and within breeding groups, examine population structure, and contextualize these findings relative to the global upland cotton gene pool. The CottonSNP63K array was used to identify 17,441 polymorphic markers in a panel of 114 diverse Pee Dee genotypes. A subset of 4,597 markers was selected to decrease marker density bias. Identity by state (IBS) pairwise distance varied substantially, ranging from 0.55 to 0.97. Pedigree-based estimates of relatedness were not very predictive of observed genetic similarities. Few rare alleles were present, with 99.1% of SNP alleles appearing within the first four breeding cycles. Population structure analysis with principal component analysis, discriminant analysis of principal components, fastSTRUCTURE, and a phylogenetic approach revealed an admixed population with moderate substructure. A small core collection (n < 20) captured 99% of the program’s allelic diversity. Allele frequency analysis indicated potential selection signatures associated with stress resistance and fiber cell growth. The results of this study will steer future utilization of the program’s germplasm resources and aid in combining program-specific beneficial alleles and maintaining genetic diversity.

Outlook for Implementation of Genomics-Based Selection in Public Cotton Breeding Programs

“…Other smaller blocks (2–5 lines) along the diagonal correspond to parent–child relationships. Previous analysis [ 17 ], together with the GRM here, show that this panel is fairly weakly structured overall, indicating the suitability of this dataset for a genome-wide association study, as long as the existing population structure is accounted for when estimating allele effects and p -values.…”

“…A total of 15,177 SNPs remained after filtering to retain only polymorphic SNPs from the CottonSNP63K array. Based on historical pedigree records, this set of 80 lines is related mostly within four or so cycles of breeding, with dozens of rounds of successive recombination resulting in admixture between groups [ 17 ]. Based on the investigation of relationships from the SNP data, it does appear that recombination has broken relationships, with the most significant exceptions occurring within two big genotype blocks, where most individuals in each of the blocks of the heatmap are (approximately) related at the full- or half-sib level ( Figure 1 ).…”

“…Eighty historical cotton lines from the PD breeding program were genotyped with the CottonSNP63K array [ 31 ], as reported in Billings et al (2021) [ 17 ]. Field trial data corresponding to these eighty cotton lines was collected between 2004–2006 at six locations across the mid-south and southeast US cotton belt [ 32 ].…”

Development of High-Yielding Upland Cotton Genotypes with Reduced Regrowth after Defoliation Using a Combination of Molecular and Conventional Approaches

“…Earlier studies conducted on 260 commercial upland cotton cultivars, released between 1970 and 1990, to determine the coefficient of parentage and pedigrees showed a narrow genetic base [3,6,13,14]. Further investigations using molecular markers also confirmed the low genetic diversity in cultivated cotton germplasm [12,[15][16][17][18][19][20].…”

Identification of quantitative trait loci for fiber quality, yield, and plant height traits in Upland cotton

“…At present, two SNP arrays are available in cotton, a 63K SNP array and an 80K SNP array (Cai et al., 2017; Hulse‐Kemp et al., 2015b). These SNP arrays have been used to assess genetic diversity (Billings et al., 2021), perform genome‐wide association studies (GWAS) (Cai et al., 2017; Gowda et al., 2022; Song et al., 2021; Sun et al., 2017), construct genetic maps (Hulse‐Kemp et al., 2015b; Zhang et al., 2019; Zhu et al., 2021), and detect QTL for fiber quality traits, yield traits, and morphological traits (Li et al., 2016; Zhang et al., 2019; Zhu et al., 2021). However, only a few studies have reported their use in identifying novel and stable QTL and develop functional markers for fiber quality traits, yield traits, and plant height in intraspecific Upland cotton populations.…”