Genetic Analysis of Flooding Tolerance in an Andean Diversity Panel of Dry Bean (Phaseolus vulgaris L.)

Abstract

Climate change models predict temporal and spatial shifts in precipitation resulting in more frequent incidents of flooding, particularly in regions with poor soil drainage. In these flooding conditions, crop losses are inevitable due to exposure of plants to hypoxia and the spread of root rot diseases. Improving the tolerance of bean cultivars to flooding is crucial to minimize crop losses. In this experiment, we evaluated the phenotypic responses of 277 genotypes from the Andean Diversity Panel to flooding at germination and seedling stages. A randomized complete block design, with a split plot arrangement, was employed for phenotyping germination rate, total weight, shoot weight, root weight, hypocotyl length, SPAD index, adventitious root rate, and survival score. A subset of genotypes (n = 20) were further evaluated under field conditions to assess correlations between field and greenhouse data and to identify the most tolerant genotypes. A genome-wide association study (GWAS) was performed using ~203 K SNP markers to understand the genetic architecture of flooding tolerance in this panel. Survival scores between field and greenhouse data were significantly correlated (r = 0.55, P = 0.01). Subsequently, a subset of the most tolerant and susceptible genotypes were evaluated under pathogenic Pythium spp. pressure. This experiment revealed a potential link between flooding tolerance and Pythium spp. resistance. Several tolerant genotypes were identified that could be used as donor parents in breeding pipelines, especially ADP-429 and ADP-604. Based on the population structure analysis, a subpopulation consisting of 20 genotypes from the Middle American gene pool was detected that also possessed the highest root weight, hypocotyl length, and adventitious root development under flooding conditions. Genomic regions associated with flooding tolerance were identified including a region on Pv08/3.2 Mb, which is associated with germination rate and resides in vicinity of SnRK1.1, a central gene involved in response of plants to hypoxia. Furthermore, a QTL at Pv07/4.7 Mb was detected that controls survival score of seedlings under flooding conditions. The association of these QTL with the survivability traits including germination rate and survival score, indicates that these loci can be used in marker-assisted selection breeding to improve flooding tolerance in the Andean germplasm.

Keywords: GWAS; abiotic stress; anoxia; common bean; flooding; waterlogging.

Figure 1

Bean plots represent the distribution of traits in non-flooded (black) and flooded (gray) conditions. The units of y-axis for each trait was mentioned after the name of the trait on the top of each graph. The arrows indicate the values of tolerant check (Royalty) for each condition. The dotted lines indicates the total means under both flooding and control conditions. No adventitious root or survival reduction was observed in control condition.

Figure 2

Population classification using a subset of 3,668 markers that had LD r² ≤ 0.1 (markers in linkage equilibrium). (A) Population structure from K = 2 to K = 4. K indicates the number of subpopulations. NA, North America; MA, Middle America (B) phylogenetic tree of genotypes, constructed by using SNPhylo. The colors of subpopulations were the same as population structure plot (K = 4).

Figure 3

Flooding indices of different subpopulations. Groups were color-coded based on the results of population structure plot (K = 4, Figure 2). Boxplots indicated by the same letter for each trait are not significantly different at 0.05 probability level.

Figure 4

Relationship among traits. (A) Pearson correlation coefficients among traits measured under flooded and non-flooded. Correlation coefficients multiplied by 100. (B) Biplot representing the relationship among different flooding indices used to separate genotypes. Traits evaluated in non-flooded and flooded conditions represented by _N and _F, respectively after the name of the trait. TW, total weight; SW, shoot weight; RW, root weight; HL, hypocotyl length; SI, SPAD index; GR, germination rate; AR, adventitious root formation; SS, survival score and SDWT, seed weight.

Figure 5

Manhattan and corresponding Q-Q plots represents the genetic architecture of different traits under flooded and non-flooded condition. Position of loci associated with seven traits in non-flooded and flooded conditions are represented. Blue dots in Manhattan plots represent SNPs falling between 0.01 and 0.1 percentile tail of the empirical distribution of P-values. Red dots represent SNPs that pass the 0.01 percentile. The name of the model selected for each trait/ treatment was mentioned on the right side of Q-Q plots.