Rogue pod traits in Phaseolus vulgaris

Graduation date: 2015Snap beans are the vegetable form of dry bean (Phaseolus vulgaris L.) with low fiber, stringless and round, succulent pods that are eaten in the immature stage. Seed companies commit significant resources to maintaining purity and uniformity of snap bean cultivars. While some variability may be introduced through outcrossing (beans are highly self-pollinated but occasional outcrosses do occur) and mechanical mixes of seed, mutation may also play a role. For traits such as pod cross-section (round vs. oval or flat), pod wall fiber and pod suture fiber (strings), reversions from the snap bean phenotype (round, low fiber, stringless) to the dry bean phenotype (oval or flat, high fiber, stringy) may occur. The reversions occur at many times the background mutation rate and are called "rogues" in the industry. Many cultivars, both commercially released and OSU bred show rogue traits and in the OSU breeding program, lines are grown out and inspected for off types every year. Among the rogue traits, strings and oval pods are of greatest concern. The reversions of round and stringless pods are of the most economically costly phenomena. In the past century scientists have studied snap bean pod traits, including suture strings, ovals and increased pod wall fiber per se. Results reported by researchers are in conflict regarding the heritability and gene models, including specific gene(s) controlling the phenotypic traits of interests. In addition, the advent of new molecular tools\ud allows us to examine the reversion phenomenon at the molecular level. In the present study, three major approaches were used to get at the rogue question. First, we observed the inheritance of traits of interest by collecting seed of 98 off type plants representing 11 cultivars and breeding lines. These plants were self-pollinated and grown and classified for phenotype over three generations. Results obtained from this study revealed that traits were initially segregating in every case, but families became increasingly fixed over time. However, observed segregation ratios failed to fit expected segregation ratios. In addition, some families that were nearly fixed still showed a few individuals of the opposite phenotype. From the selfing of individual plants, we saw skewed segregation ratios that did not fit previous published inheritance models, any classical Mendalian one or two gene models or the populations within varieties segregating in a similar manner. Analyzing phenotype data showed an association of string with low or moderately low fiber content, which is congruent with non-rogue fixed pods, while oval pods showed association to high or moderately high fiber content. The second avenue of research was to make controlled crosses to conduct tests of allelism and inheritance. The objective of inheritance crosses was to decipher the dominance-recessive relationships of strings and ovals. The test of allelism, by crossing two stringy off type plants was to determine if strings were controlled by the same or different loci in different genetic backgrounds. A test of inheritance, by crossing an off type to a parental plant (string by stringless), was made to shed light on the genetic model of the trait. Due to time limitations, only the first hybrid generation was grown, which revealed dominance for the stringless and round pod phenotype. The third research strategy was to initiate molecular analysis using genotyping by sequencing (GBS) of a pooled sample of stringless genotypes compared to a pooled set of stringy genotypes to find single nucleotide polymorphisms (SNPs) between the two groups that would suggest candidate gene(s) for string (or lack of string) formation. Ninety-six samples split into two genotype pools (stringy and stringless), both with low fiber content were sequenced. In the stringy pool 31,212 SNPs were initially found and 38,936 SNPs were found in the stringless pool. SNPs with low quality and probability, due to low read depth, and SNPs that did not have a high impact on the gene itself (high impact SNPs are those with altered start, stop or splice sites that change gene function) were removed. Forty-three SNPs in the stringless pool that differed from the reference genome and 41 in the stringy pool were annotated as candidate genomic regions of control. Due to the fact that off types within the same family showed a lack of homogeneity for the\ud phenotypic data within varieties, the high impact polymorphisms could have been due to factors other than the trait of interest. None of our SNPs of interest were present in published candidate genes for pod suture strings, fiber, or pod shattering, but novel areas of the genome were identified. Using our SNP data and putative candidate regions that have been identified, expression in these regions can be analyzed to further refine the list of potential candidate gene(s) controlling strings