Breeding Snap Beans For Organic Agriculture: Quantification And Application Of Key Traits

Objective 1: Our first objective was to gain understanding on how population-wide differences in germination rate and seedling vigor relate to previous breeding under organic or conventional management.

To achieve this, we grew each of the four research populations in two fields, one under management conventional and one under organic management. Emergence date, seedling counts, and seedling vigor data was collected for all populations in both fields. Collected data was analyzed, and any traits of significance related to early-season performance were explored through QTL (quantitative trait loci) analysis. 

We SNP genotyped all families from the four populations and generated a set of linkage maps was generated for each population. SNP mapping, or single-nucleotide polymorphism mapping, involves identifying genetic markers (polymorphisms) within all of the individuals of a population. This allows for comparison of phenotypes across a large population, and the determination of genetic regions that are associated with certain phenotypes. Our new maps provided remarkably more markers compared to the original maps generated for these populations, shown in Table 1. The new ORBV maps are still under development but are expected to show similar improvements in market density and coverage.

Table 1: Numbers of markers pre-and post-JoinMap, individuals used in map after removing heterozygous and identical individuals, and overall cM (centimorgans) of each map (modified from (King, 2019).

^yBased on 1,200 cM (centimorgan) average genetic map length.

Germination data were explored for significance using ANOVAs. Significance tests were run on the F₇ and F₈ generations of the four populations to explore the main effects (i.e., seed color, cotyledon color, breeding history, cross, production system, and production system by breeding history interactions), shown in Table 2.

Table 2: Mean squares and degrees of freedom from ANOVA on early-season phenotypic data from F7 and F8 generations of four recombinant-inbred line snap bean populations, from data collected in 2020 and 2021 at the Vegetable Research Farm in Corvallis, OR.

Significance codes: * indicates significance at <0.05; ** indicates significance at <0.01; *** indicates significance at <0.001.

We then utilized the data collected on the F₇ and F­₈ generations, as well as F₆ data from that was previously taken on the same populations to explore relationships between the main effects. Figure 1 depicts germination percentage in relation to other main effects of interest.

Figure 1: Germination percentage of all families from all populations within various treatments, collected from the F6-F8 generations of snap bean populations grown in conventional and organic production systems at the OSU Vegetable Research Farm in 2020 and 2021.

We also sought to understand how seed color related to germination percentage specifically, so a corresponding figure was produced using data from the white-seeded families within each population. Within the white-seeded families, organically bred populations had higher germination in the 2020 organic environment as well as both 2021 environments, as described in Figure 2. In 2018 and the 2020 conventional environment the conventionally bred populations had higher germination. This suggests that white-seeded families from an organic background were performing better than their conventional counterparts, unlike the trends we saw in the populations overall.

Figure 2: Germination percentage of white-seeded families from all populations within various treatments, collected from the F6-F8 generations of snap bean populations grown in conventional and organic production systems at the OSU Vegetable Research Farm in 2020 and 2021.

We compared the days to germination for the same data within all families of all populations and the white-seeded families alone. For both, a significantly faster rate of germination was observed in organic populations, at the p<0.001 level. In terms of scale, the difference between the white-seeded families of varying breeding histories was twice as large as the difference between all families of the same breeding histories. This comparison can be seen between Figure 3 and Figure 4.

^z For each breeding history group, mean days to germination for each population is represented by a dotted vertical line. Color is shown according to breeding history, or as a blended color where the density values overlap.

Figure 3: Days to germination of all families from all populations within various treatments, collected from the F6-F8 generations of snap bean populations grown in conventional and organic production systems at the OSU Vegetable Research Farm.

^z For each breeding history group, mean days to germination for each population is represented by a dotted vertical line. Color is shown according to breeding history, or as a blended color where the density values overlap.

Figure 4: Days to germination of white-seeded families from all populations within various treatments, collected from the F6-F8 generations of snap bean populations grown in conventional and organic production systems at the OSU Vegetable Research Farm.

This reduction in days to germination was observed within organically bred, white-seeded families suggests possible adaptation to reduce the risk of pathogen exposure during pre-emergence. Heritability estimates for these traits, shown in Table 3, indicate moderate potential for pursuing improvement of these traits via plant breeding.

Table 3: Narrow-sense heritability estimates (h²±SE) for early-season performance traits in the F7 and F8 generations of snap bean populations, derived from ANOVA tables generated from data collected on replicated trials, separated by production system.

Using the improved linkage maps, QTL were identified related to early season performance traits within each of the four populations. 

In the ORBV-O population many QTL related to early season performance were identified on chromosomes 7 and 8. Four QTL were identified in the ORBV-C population related to early season performance.

In the HYPR-O population, QTL associated with days to germination were identified on Pv02 and Pv08. These were observed across multiple years in the organic production system. The additive effect for each of these was negative, indicative of faster germination when an allele of Provider is substituted for Hystyle. QTL related to germination percentage and vigor were also identified on Pv02 and Pv08, as well as on Pv10. These had predominantly positive additive effects, indicative of higher germination and vigor when an allele of Provider is substituted for Hystyle.

In the HYPR-C population, QTL related to days to germination were identified on linkage groups Pv01, Pv02, Pv07, Pv08, Pv10, and Pv11. QTL identified on Pv01, Pv02, Pv07, and Pv11 had positive additive effects, indicating longer days to germination associated with an allele of Provider substituted for that of Hystyle. These trends were observed across multiple years and production systems. QTL associated with germination percentage and vigor were located primarily on linkage groups Pv02, 08, and 10, and had positive additive effects across years and production systems.

Objective 2:

Our second objective was to quantify differences in functionality of plant-rhizobia interactions within and between populations with comparisons made to the breeding histories of each population.

To achieve this, we explored root architecture in a field environment, followed by greenhouse studies. Our work revealed notable changes across production systems, indicative of altered resource-seeking behaviors between the two environments. Notably, the roots within the organic productions system had shallower taproot angles, and larger basal root diameters. These changes may be explained by plastic phenotypic responses to phosphorous availability, as well as an increased mycorrhizal dependency. Previous findings have indicated a reduced orientation, or shallower root angle associated with reduced phosphorous availability in the soil (Bonser et al., 1996). Further, thicker roots may have reduced water and nutrient uptake, but result in greater root longevity and increased mycorrhizal relationships (Eissenstat, 1992).

Contrary to expectations the organically bred populations had reduced nodule numbers in comparison with their conventional counterparts, however these observations do not account for variation in functionality of the nodules. Given evidence of loss of specificity in conventionally bred legumes, the higher number of nodules may actually be a factor of heightened promiscuity in nodulation as opposed to an increase in functional symbiotic relationships (Kiers et al., 2007).

Narrow-sense heritability estimates generated for the root morphological data were highly variable both across and within the two production environments. Most traits had negative or very low heritability, indicative of the predominant role of the soil environment in shaping root architecture.

Loss of specificity in root-associated nodules in the conventionally bred populations of snap beans is the predominant question remaining following our inquiry and analysis of root morphology.

Objective 3: Our third objective was to characterize the root microbiome within a subset of each of the four populations using 16S and ITS sequencing. 

Sequence analysis of the 16s rRNA gene from rhizosphere soil samples provided evidence of a non-significant trend towards increases in within-group microbiome diversity tied to breeding history. Comparing between-group diversity, however, revealed significant shifts in the microbial community in samples from either organic or conventional breeding histories. These shifts are visually represented in Figure 5 and Figure 6.

Figure 5: Canonical analysis of principal components ordination of snap bean rhizosphere samples based on breeding history, generated from 16s sequences, with ANOVA of the model shown in the table.

Figure 6: Canonical analysis of principal components ordination of snap bean rhizosphere samples based on breeding history, generated from ITS sequences, with ANOVA of the model shown in the table.

Exploring these significant relationships further, a sPLS-DA (sparse Partial Least Squares for Discriminant Analysis) model was tuned to fit the data described. Using the classification that was developed through this model tuning, it was possible to estimate the predicted ordination area for each class, shown in Figure 7 and Figure 8.

Figure 7: Background prediction areas shown with sample ordinations from snap beans of variable breeding histories, as calculated from maximum distance methods within the sPLS-DA for 16s data of rhizosphere bacteria in organically managed plots at the Lewis Brown Research Farm.

Figure 8: Background prediction areas shown with sample ordinations from snap beans of variable breeding histories, as calculated from maximum distance methods within the sPLS-DA for ITS data of rhizosphere fungi in organically managed plots at the Lewis Brown Research Farm.

From the ordination loadings it was possible to identify the sequence information for the taxa that were most responsible for shifts between each of the two classes in each data set. Changes in the mycobiome between groups revealed the communities to be significantly shifted by both breeding history and parentage. Both pathogenic and potentially beneficial fungi were identified in the rhizosphere communities of organically and conventionally bred beans. Among the beneficial taxa with increased abundance in the organically bred bean rhizospheres were an ectomycorrhizal species and an arbuscular mycorrhizal species. Although functionally quite different, both of these fungal symbionts are known to aid host plants in nutrient acquisition, disease resistance, and abiotic stress tolerance (Parniske, 2008; Lehmann et al., 2012; Hohmann & Messmer, 2017).

Objective 4: Our fourth objective was to utilize the existing four populations to develop a new snap bean variety in collaboration with a regional seed company and local vegetable farmers.

Using data from two field seasons at two locations (a university research farm and a collaborator’s farm), along with data from a taste test with 20 participants, we were able to identify 3 candidate lines for release. The attached Participatory Plant Breeding PowerPoint discusses that status and future steps of these lines.

Sources Cited:

Bonser, A. M., Lynch, J., & Snapp, S. (1996). Effect of phosphorus deficiency on growth angle of basal roots in Phaseolus vulgaris. New Phytologist, 132(2), 281–288. https://doi.org/10.1111/j.1469-8137.1996.tb01847.x

Eissenstat, D. M. (1992). Costs and benefits of constructing roots of small diameter. Journal of Plant Nutrition, 15(6–7), 763–782. https://doi.org/10.1080/01904169209364361

Hohmann, P., & Messmer, M. M. (2017). Breeding for mycorrhizal symbiosis: Focus on disease resistance. Euphytica, 213(5), 113. https://doi.org/10.1007/s10681-017-1900-x

Kiers, E. T., Hutton, M. G., & Denison, R. F. (2007). Human selection and the relaxation of legume defences against ineffective rhizobia. Proceedings of the Royal Society B: Biological Sciences, 274(1629), 3119–3126. https://doi.org/10.1098/rspb.2007.1187

King, R. M. (2019). Breeding Snap Beans for Organic Agriculture: Genomic Shifts under Different Agricultural Management Systems [Masters Thesis, Oregon State University]. https://ir.library.oregonstate.edu/concern/graduate_thesis_or_dissertations/kp78gn560

Lehmann, A., Barto, E. K., Powell, J. R., & Rillig, M. C. (2012). Mycorrhizal responsiveness trends in annual crop plants and their wild relatives—A meta-analysis on studies from 1981 to 2010. Plant and Soil, 355(1–2), 231–250. https://doi.org/10.1007/s11104-011-1095-1

Parniske, M. (2008). Arbuscular mycorrhiza: The mother of plant root endosymbioses. Nature Reviews Microbiology, 6(10), 763–775. https://doi.org/10.1038/nrmicro1987