Identification of favorable haplotypes/alleles and candidate genes for three plant architecture-related traits via a restricted two-stage multilocus genome-wide association study in upland cotton

Plant architecture-related traits are extremely crucial for cotton plants and are closely linked to their ability to adapt to reasonably close planting and to their suitability for mechanized harvesting. However, the genetic basis for cotton plant architecture-related traits remains poorly understood. To identify favorable haplotypes/alleles (FHs/FAs) and candidate genes for target traits, we performed a restricted two-stage multilocus genome-wide association study (RTM-GWAS) using a natural population consisting of 315 upland cotton accessions and 9244 single-nucleotide polymorphism linkage disequilibrium blocks (SNPLDBs), which were explored and established using the specific-locus amplified fragment sequencing (SLAF-seq) technique. Totals of 68, 34 and 55 significant SNPLDB loci associated with the plant height (PH), fruit branch angle (FBA) and fruit branch length (FBL), respectively, were identified via the RTM-GWAS method. We then estimated the allele effects of the significant loci associated with the target traits and further generated SNPLDB allele matrices to provide a brief inheritance constitution of the population. Among the significant SNPLDB associations, five involved in four major loci (LDB_1_21982826, LDB_9_51578589_51578819, LDB_16_37952328 and LDB_19_52309050_52309284) for the PH or FBL were concurrently detected in a multiple-environment combined model and in at least one single-environment model. A comparison of FHs/FAs frequencies for the four major SNPLDB loci revealed that the frequencies of FHs or FAs that were hidden in short and compact cotton accessions were much higher than those in tall and noncompact cotton accessions. We also observed that cotton accessions from the northern super-early-maturation region (NSER) had relatively short and compact plants compared with those from the Yangtze River region (YZRR) and northwestern inland region (NIR), and the high FA frequencies of the four major SNPLDB loci were consistent with the relatively short and compact plant architecture in five different geographic regions. In the neighboring genome fragments of the four major SNPLDB loci, ultimately, the five candidate genes that were highly expressed during the early stages of the growth and were responsible for cotton PH and FBL through the auxin signaling pathway were explored by RNA-seq data and bioinformatics analysis. This study identified FHs/FAs and candidate genes for PH and FBL, which could provide key insights into the genetic basis for the plant architecture of upland cotton and could promote the breeding of new cotton varieties with ideal plant architecture suitable for mechanized harvesting.