|12-184 Project Manager: D. C. Jones|
BRIDGING GENOMICS AND PLANT BREEDING WITH A COTTON NESTED ASSOCIATION MAPPING POPULATION
Michael A. Gore, USDA-ARS
The primary objectives of this 2012 study were to (i) select 20 diverse founders for constructing a cotton NAM population of 3,000 F5 recombinant inbred lines, and (ii) evaluate a cotton diversity panel of 384 lines for lint yield and fiber quality traits.
We constructed a panel of 381 upland cotton lines to represent the genetic diversity present in public-sector cotton breeding programs across the United States. In addition, three G. barbadense lines were included in the panel to serve as an outgroup for phylogenetic analyses. These 384 lines were genotyped with 120 fluorescent-labeled microsatellite markers on an ABI 3730 capillary sequencer. Notably, these 384 lines were selfed for several generations to increase their homozygosity and phenotypic uniformity. Indeed, these 384 lines were highly homozygous (average = 2.2%) and captured a total of 556 microsatellite alleles. There was an average of 4.63 alleles per microsatellite marker, and the average polymorphism information content (PIC) value (i.e., relative informativeness) per marker was 0.16. Following expectations for population genetics theory, the percentage of rare alleles (alleles only found in 0.5% accessions) in the examined upland cotton germplasm was found to be 21% and most of the alleles had a very low minor allele frequency. Analysis of population structure using the software STRUCTURE identified four discrete subpopulations and admixture. Similarly, we also identified four major clusters for the upland cotton germplasm with a neighbor-joining analysis. The identified four subpopulations are largely concordant with the geographical distribution of US cotton breeding programs.
The 20 NAM parents will be: PI 529169 ACALA 5, PI 529318 ALLEN 33, PI603003 CD3HCABCUH-1-89, PI 528970 DELTAPINE 14, PI 529522 DES 24, PI 529021 DIXIE KING, PI529572 FJA, PI 528560 M.U.8B UA 7-44, PI 583376 NC 88-95, PI606814 PAYMASTER HS26, PI529617 PD 2164, PI 533644 PD 785, PI607237 SA 2413, PI 528872 SEALAND #2 (G.B. X G.H.), PI601652 SOUTHLAND M1, PI 528965 STATION MILLER, PI 529447 TASHKENT 1, PI 528839 TIDEWATER 29 (G.B. X G.H.), PI 528963 TOOLE, and SA-0296 WANNAMAKER CLEVELAND. To select parents for constructing an elite NAM population, we assembled 10 core sets from the cotton diversity panel with sizes ranging from 8 to 53 lines in increments of 5 lines by maximizing allelic richness with a simulated annealing algorithm in the software POWERMARKER. Before constructing the 10 core sets, we eliminated lines that had an unfavorable mutant phenotype, colored lint, late maturity, or higher than 1% G. barbadense introgression. This resulted in a final sample size of 324 upland lines that captured 322 alleles for the selection of NAM parent lines. The constructed core sets were constrained to always include the following six upland cotton lines: ACALA MAXXA (Cluster I representative/Paterson NAM parent), DELTAPINE 14 (Cluster IV representative), DES 56 (common parent/Paterson NAM parent), DIXIE KING (Cluster III representative), PAYMASTER HS200 (Paterson NAM parent), and WANNAMAKER CLEVELAND (Cluster II representative). To strike a balance between allelic richness and the costs of population construction, we selected 20 diverse lines to cross separately as males to the common parent DES 56. These 20 diverse lines plus DES 56, ACALA MAXXA, and PAYMASTER HS200, capture 84% of the 322 alleles identified in the 324 upland cotton lines, while 74% of the total alleles identified in the complete set of 381 upland cotton lines.
To complement the NAM population, we evaluated 348 of the 384 lines that had sufficient quantities of seed for planting at four field locations (AZ, NC, SC, and AR). Within a location, the experimental design was a 16 × 29 incomplete block α-lattice design. The α-design was augmented by including four different commercial varieties within each incomplete block. The order of the 16 entries in each incomplete block was randomized. Because of the large size of the experiment, only a single replicate was grown at each of the locations. Experimental units were one-row 10.67-m plots. Plots of a conventional cotton cultivar were planted as border on all sides of the experiment. All plots were grown with standard cultivation and pest control methods. All plots were machine harvested for measurement of lint yield. Prior to mechanical harvest, 25-boll samples were harvested by hand. The collected seedcotton samples are in the process of being ginned to allow for the measurement of boll and fiber quality traits. These phenotypic data will be eventually combined with genotypic data for genome-wide association and prediction studies.
|Project Year: 2012|
Search 2012 Projects:|
▸ New Mexico
▸ North Carolina
▸ South Carolina
▸ Cotton Incorporated Fellow
▸ Crop Improvement
▸ Production Efficiency
▸ Sustainable Cotton
▸ Variety Improvement
▸ All Project Nos.