A Comparison of Publicly Available QTL SSRS for MAS with Traditional Plant Breeding Selection Methods in Cotton

Abstract

Public breeding programs for upland cotton (Gossypium hirsutum) underutilize genotypic selection methods and specifically marker assisted selection for fiber quality selection. A previous study by Kari Hugie sought to analyze publicly available simple sequence repeat (SSR) markers in three diverse populations to quantify those that showed correlative trait stability in different backgrounds. Stable markers identified from that study, 6 for fiber bundle strength (FBS) and 6 for upper half mean length (UHML), as well as two additional markers identified by Dr. Fang of the USDA Louisiana Laboratory were then utilized to make selections in two G. hirsutum populations. Population 1 (TAM 11K-13 ELSU/Del Cerro//13P-54 ELSU) was of interspecific background and Population 2 (TAM 11K-13 ELSU/TAM06WE-621 ESU) was of intraspecific background. In 2016, both populations were grown at the Agrilife Research Center in College Station, Tx and individual plant selections were made separately on the basis of SSR marker number and phenotype for both FBS and UHML to form populations divergent for the trait of interest. Selections were planted in progeny rows in 2017 and fiber samples collected for each row. Fiber data was analyzed to compare selection efficacy within the divergently selected subpopulations and between the marker and phenotypically selected populations. Divergent selections made on the basis of number of markers in the desired allelic state failed to show significant differences between subpopulations while those made on the basis of high and low value for the phenotypic trait were significant for both traits in both populations. Analyzing individual markers, none were found to be significant (p < 0.05) for either trait in both populations for both years. In Population 1 for UHML, two markers were found to be significant only in 2016, CIR196v197 (p=0.0416) and NAU1369v247 (p=.0011), and for FBS in 2016 only CGR6329v232 was significant (p= 0.0278). In Population 2, marker BNL1604v98 was significant in both years for UHML (p=0.0270 and p=.0448 in 2016 and 2017 respectively), and in 2016 for FBS (p=0.0425). Despite significance, numerical differences between selections with or without the BNL1604v98 allele were small.