The GAMA Approach for the Analysis of Large Germplasm Collections: Genetic Diversity and Landscape Genomics of Common Bean (Phaseolus Vulgaris) Landraces in Brazil

doi:10.21203/rs.3.rs-1310635/v1

Download PDF

Research Article

The GAMA Approach for the Analysis of Large Germplasm Collections: Genetic Diversity and Landscape Genomics of Common Bean (Phaseolus Vulgaris) Landraces in Brazil

https://doi.org/10.21203/rs.3.rs-1310635/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Common bean (Phaseolus vulgaris L.) is an essential staple crop in the Brazilian diet, being the major source of dietary protein in this tropical country. In addition, Brazil is also one the biggest producer and consumer of this crop worldwide. Brazil covers a wide and diverse geographical area in South America, where common bean is cultivated both in tropical climates in the North and Central regions, or sub-tropical ones in the South. This climatic diversity in Brazil is reflected in the wide genetic diversity of Brazilian common bean landraces, which could represent an important resource for breeding improvement of this crop. In the current study, we characterized the genetic diversity, population structure and linkage disequilibrium in a collection of 274 georeferenced Brazilian landraces using Single Nucleotide Polymorphisms obtained with Genotype-By-Sequencing. We integrated geographic information system, climatic conditions, agronomic (market class information), and molecular diversity analyses, labeled here GAMA (Geographic Agronomic Molecular Approach), to identify patterns of selection in this collection. This combined landscape genomic analysis allowed the identification of candidate genes or genomic regions related to abiotic stress response in this species, but also possible donor accessions for drought and heat tolerance in this collection. This information can facilitate the pre-selection of smaller samples of Brazilian landraces diversity and identify candidate markers and genes for in-depth evaluation in breeding programs aimed at the improvement of this crop.

Brazil is one of the largest producers of common bean (Phaseolus vulgaris) worldwide, with more than 3 M metric tons of dry beans produced each year. It is also one of the main consumers of this crop (Petry et al. 2015). Indeed, common bean represents an essential staple crop in the Brazilian diet, contributing most of the protein intake for most of the population in this country (da Silva and Wander 2013). Although common bean is not a native crop of Brazil, it has been cultivated in this area by native populations for a long time before the arrival of the Europeans, according to both archaeological and historical reports (Hoehne 1937; Freitas 2006). Nowadays, several different local landraces are cultivated in Brazil, principally by smallholder-farmers who rely on rainfall for the irrigation of this crop (Vieira 1988, 1998; Barbosa and Gonzaga 2012). Common bean is cultivated in a wide variety of environments, from a tropical climate in the Northern and Central states to a sub-tropical climate in the South of Brazil, using different cropping systems (Melo 2009; Barbosa and Gonzaga 2012; da Silva and Wander 2013). The wide diversity of Brazilian landraces constitutes an important potential source of tolerance to both abiotic and biotic stresses for improving this species and developing locally adapted elite cultivars specific for this country (Burle et al. 2010).

Common bean is a diploid plant (2n = 22) with an estimated genome size of 580Mb that has been sequenced recently (Andean gene pool: G19833 accession, Schmutz et al. 2014; Mesoamerican gene pool: breeding line BAT93, Vlasova et al. 2016; Rendón-Anaya et al. 2017). The genus Phaseolus originated in Mesoamerica where most of the species of this genus are distributed (Freytag and Debouck 2002). Wild common bean was dispersed through several discrete long-distance events from Mesoamerica to Central and South America resulting in three major, genetically and geographically separated, wild gene pools named Mesoamerican, Ancestral, and Southern Andean gene pools (Debouck et al. 1993; Gepts 1998; Bitocchi et al. 2012; Rodriguez et al. 2016; Ariani et al. 2018). Interestingly, no wild common bean has been identified in Brazil to date. Of the three wild gene pool of this species, two of them have been independently domesticated about 10,000 years ago in what is now Mexico and Argentina. This double domestication generated the Mesoamerican and Andean domesticated gene pools of this species (Gepts et al. 1986; Kwak and Gepts 2009; Mamidi et al. 2011, 2013). The two domesticated gene pools can be differentiated either using morphological characteristics, or a wide array of molecular and protein markers or sequences (Gepts et al. 1986; Koenig and Gepts 1989; Chacón et al. 2005; Kwak and Gepts 2009; Mamidi et al. 2011, 2013; Bitocchi et al. 2013). In addition, these two gene pools are separated by partial reproductive isolation in both wild and domesticated populations, which causes hybrid weakness in F₁ and later generations (Gepts and Bliss 1985; Koinange and Gepts 1992; Hannah et al. 2007).

Previous studies about genetic diversity in Brazilian landraces identified the presence of both Mesoamerican and Andean gene pools in Brazil based on the study of phaseolin seed proteins. In particular, most of the landraces analyzed carried a phaseolin type characteristic of the Mesoamerican gene pool (‘S’ type), while other showed the phaseolin specific of the Andean group (‘T’ type) (Gepts et al. 1988). Further studies using microsatellite markers confirmed the differentiation between Mesoamerican and Andean gene pool in Brazilian landraces, but also identified several hybrids between the Mesoamerican groups, suggesting a certain degree of genetic admixture in common bean landraces from Brazil (Burle et al. 2010). A similar pattern of genetic admixture within the Mesoamerican gene pool of Brazilian landraces has been recently observed using Next-Generation-Sequencing-based markers like DArT-Seq SNPs (Valdisser et al. 2017).

One of the principal difficulties in exploiting large genetic resource collections, i.e., “gene banks,” for crop genetic improvement is the general inability of genetically characterizing these collections because of the difficulty or cumbersomeness of accurately phenotyping traits of agronomic interest. For example, an assessment of drought tolerance is cumbersome because it requires a comparison of yields measured under well-watered and various drought stress treatments for each accession of the collection. Here, we propose a strategy – called GAMA (Geographic, Agronomic, Molecular Approach) - to narrow down the number of accessions to be preferentially screened for a phenotypic trait, such as drought tolerance. This strategy includes, as a condition sine qua non, a geographic activity in which the coordinates of accessions are used to correlate their location of origin with relevant climatic variables, such as precipitation, temperature, or potential evapotranspiration. Thus, a subset of accessions can be identified that originate, for example, in more highly drought-stressed regions, an approach that has been called Focused Identification of Germplasm Strategy (FIGS; Khazaei et al. 2013). It remains to be tested whether these accessions are indeed better sources of genes for stress tolerance.

The integration of environmental and genetic data, also known as landscape genomics (Schoville et al. 2012; Bragg et al. 2015), could facilitate this process by allowing a pre-selection of a smaller germplasm sample and the identification of candidate markers for in-depth future evaluation and gene identification. Landscape genomics has been applied mostly for identifying signatures of environmental adaptation in wild natural populations (Schoville et al. 2012; Bragg et al. 2015) with the aim of guiding improvement of crop plants (Henry and Nevo 2014). However, by integrating information about agronomic practices and climatic data during the crop-growing season, this approach has been applied also to local landraces of crop plants like maize and barley (Westengen et al. 2012; Abebe et al. 2015).

A second activity in the GAMA approach includes the agronomic evaluation of the accessions identified in the first activity for relevant phenotypes potentially involved in drought stress tolerance (e.g., Beebe et al. 2008, 2010, 2013; Berny Mier y Teran et al. 2019a,b; Parker et al. 2020). Finally, in the third activity molecular polymorphisms are identified that correlate with the environmental variables chosen in the first activity and, if available, the results of the phenotyping in the second activity. Thus, the GAMA approach illustrated here can identify and target a subsample of a gene bank through geographic information system analysis and concurrently characterize this subsample both phenotypically and genotypically.

In the current study, we characterized the genetic diversity of a collection of 274 local landraces from Brazil representative of the diversity and geographical distribution of this crop at approximately 13,000 SNPs using Genotyping-By-Sequencing (GBS) technology (Elshire et al. 2011). By integrating this SNPs dataset with market class information and climatic data estimated for the growing season in different areas of Brazil, we were able to identify possible donor of drought and heat tolerance in this collection, but also putative markers that could be used for breeding improvement of Brazilian bean germplasm. This information would be extremely useful for exploiting the genetic diversity available in Brazilian common-bean landraces, especially in breeding for resilience to climate change.

Genotypic data and diversity analysis.

After imputation and final filtering, we retained 259 individuals genotyped at 13,220 loci. SNPs were generally evenly distributed across the 11 chromosomes, ranging from 818 SNPs on chromosome Pv10 to 1,583 SNPs on chromosome Pv02 (Table 1). A plot of genome-wide π showed the highest nucleotide diversity in some genomic fragments on chromosomes Pv01, Pv02, Pv05, Pv06, and Pv11 (Fig. 1A), even though the largest number of sliding windows with π higher than the 99th percentile of its distribution was located on chromosomes Pv01 and Pv09 (Fig. 1A, Table 1). Chromosome Pv10 had the lowest average π (1.41e^− 5), while chromosome Pv01 had the highest nucleotide diversity (1.84e^− 5) (Table 1); however, there were variations between the Andean, Mesoamerican, and hybrid gene pools in this sample (Fig. S1). A plot of Minor Allele Frequency (MAF) distribution showed that most of the identified SNPs had a MAF around 0.2 (Fig. 1B), with 8% of the markers having a MAF ≤ 0.1.

Table 1

Distribution of SNPs and sequence diversity across chromosomes of common bean in a Brazilian landrace sample
Chromosome	Total SNPs	π	High π windows	Half-decay distance r²v (bp)
Pv01	1,515	1.84E-05	23	1,882,828
Pv02	1,583	1.64E-05	14	3,277,857
Pv03	1,413	1.71E-05	7	6,223,743
Pv04	994	1.52E-05	0	1,665,838
Pv05	1,101	1.59E-05	7	3,164,257
Pv06	978	1.63E-05	7	2,576,845
Pv07	1,040	1.55E-05	0	3,098,186
Pv08	1,253	1.44E-05	3	1,252,883
Pv09	1,200	1.73E-05	22	12,640,750
Pv10	818	1.41E-05	0	7,695,267
Pv11	1,325	1.61E-05	12	13,666,074

By dividing the genotypes analyzed according to population structure analysis and filtering out the SNPs with more than 20% missing data in each group, a similar number of SNPs was retained, with 12,563 SNPs for ‘pure Mesoamerican’ (M), 12,202 SNPs for ‘pure Andean’ (A), and 13,049 SNPs for ‘Mesoamerican hybrids’ (MH). These groups showed different patterns of π across common bean chromosomes (Fig. 1). In particular, the ‘pure Andean’ group showed the highest relative diversity on chromosomes 1 and 3, while the ‘Mesoamerican hybrids’ had higher diversity than the ‘pure Mesoamerican’ on many of the chromosomes.

Population structure, genetic PCA, and sPCA.

Population clustering analysis using two sub-populations (K = 2) clearly distinguished between the Mesoamerican and Andean gene pools of this species, with most of the accessions belonging to the Mesoamerican gene pool (Fig. 2A). More specifically, 204 accessions clustered in the Mesoamerican gene pool, 51 in the Andean, and 4 were identified as hybrids between these two gene pools (membership coefficient < 0.7; Table S2). This population differentiation was confirmed by genetic PCA, where PC1 clearly discriminated between Andean and Mesoamerican genotypes (Fig. 2B). Plot of cross-entropy values for different number of sub-populations identified K = 5 as the best number of sub-populations for describing our dataset (Fig. S2). Population clustering showed that the Andean genotypes (as inferred by K = 2) formed a separate sub-population, while the Mesoamerican group included four “pure” (or non-hybrids) sub-populations combined with many intermixed individuals (Fig. 3A). Most of the Mesoamerican genotypes analyzed belonged to the M2 sub-population (67 accession); six individuals each belonged to the M1 and M3 sub-populations, 11 to the M4 sub-population, and 114 were identified as hybrids among Mesoamerican sub-populations (membership coefficient < 0.70; Table S2). Genetic PCA showed that the Mesoamerican sub-populations were differentiated mainly across PC2 and PC3, with many hybrid individuals across these two axes (Fig. 3B).

sPCA analysis identified a strong local structure in the collection analyzed (Fig. 4A), suggesting a negative spatial autocorrelation with high genetic diversity between physically close genotypes (based on geographic coordinates). Mantel’s test for correlation between geographic and genetic distances showed no correlation between these two variables (r = ~-0.04, P-value = 0.98), with the observed r falling within a simulated distribution with no spatial genetic structure (r = 0, i.e., null hypothesis) (Fig. 4B).

Kinship and linkage disequilibrium analysis.

Kinship analysis clearly showed the presence of two main groups in the sample, representing the Mesoamerican and Andean gene pools of common bean (Fig. S3). The estimated kinship and population clustering matrix were used for correcting LD by relatedness (kinship) and population structure in the sample (Mangin et al. 2012). This method was able to correct to a large extent for inflated LD due to kinship and structure (Fig. 5). Visual examination of LD heatmap across chromosomes showed that correction by kinship (r²_v) outperformed the correction by population structure (r²_s), while the correction using both kinship and population structure (r²_vs) was in general the best one at correcting inflated LD. A similar pattern could be observed in plots of LD decay across the 11 chromosomes of common bean (Fig. 6), even though the LD curves showed different decays among chromosomes. Overall, LD reached half-decay within 1.2 to 3.3Mb (Fig. 6, Table 1). However, chromosomes Pv03, Pv09, Pv10, and Pv11 showed a slower LD decay, with chromosome Pv09 and Pv11 having an estimated half-decay distance of more than 12Mb (Table 1).

Association analysis for rainfall and mean temperature estimated for the cultivation period.

Rainfall and mean temperature showed similar distribution during the warmest and wettest quarters (Fig. 7A, B). However, average rainfall was generally lower during the warmest quarter, and the varieties grown during this period were cultivated with just around 200 mm of rainfall (Fig. 7A). Mean temperature was slightly higher during the warmest quarter; however, the distribution of mean temperature was wider during the wettest quarter, and, in some cases, varieties grown during this wettest quarter presented mean temperatures lower than 15° C for the period (Fig. 7B). Manual evaluation of QQ-plots for GWAS analysis with temperature and rainfall during the warmest and wettest quarters showed an inflation of observed P-values when not including genetic PCA data for correcting for population structure (Fig. S4). When including PCA results for correcting for population structure in the association analysis, the QQ-plots showed a normal P-value distribution with a few strong associations (Fig. S5).

Association analysis for PrecWet identified two significantly associated SNPs on chromosome Pv02 and Pv08 of common bean reference genome (Fig. 8A), tagging 19 genes. Among these candidate genes, the most interesting ones, based on the available annotation, could be found on chromosome Pv02. Indeed, in the 100Kb window surrounding the significant SNPs on chromosome Pv02 we identified several gene models homologous to Arabidopsis, including a poly(A) polymerase 1 gene models involved in cellulose and ethylene biosynthesis, a cyclin-dependent kinase, and a gene model involved in plastid movement (Table 2).

Table 2

Candidate genes identified by genome-wide association analysis for association with PrecWet (BIO16 - Precipitation of wettest quarter)
GeneID	Arabidopsis homolog	Description
Chromosome Pv02
Phvul.002G325600	AT3G18215.1	Protein of unknown function, DUF599
Phvul.002G325700	AT3G18215.1	Protein of unknown function, DUF599
Phvul.002G325800	AT3G18215.1	Protein of unknown function, DUF599
Phvul.002G325900	PAPS1	Poly(A) polymerase 1
Phvul.002G326000	CC1	COMPANION OF CELLULOSE SYNTHASE 1
Phvul.002G326100	AT3G18210.1	2-oxoglutarate (2OG) and Fe(II)-dependent oxygenase superfamily protein
Phvul.002G326200	AT3G61520.1	Pentatricopeptide repeat (PPR) superfamily protein
Phvul.002G326300	AT3G18190.1	TCP-1/cpn60 chaperonin family protein
Phvul.002G326500	AT3G18180.1	Glycosyltransferase family 61 protein
Phvul.002G326600	ACO4, EAT1, EFE	Ethylene-forming enzyme
Phvul.002G326700	AT5G52450.1	MATE efflux family protein
Phvul.002G326800	AT;CDKD;1, CAK3AT, CDKD1;1	Cyclin-dependent kinase D1;1
Phvul.002G326900	PMI1	Plastid movement impaired1
Chromosome Pv08
Phvul.008G166100
Phvul.008G166200	AT5G54970.1
Phvul.008G166300
Phvul.008G166400	AT5G54970.1
Phvul.008G166500	AT1G77700.1	Pathogenesis-related thaumatin superfamily protein
Phvul.008G166600	AT1G38131.1	O-fucosyltransferase family protein

Association analysis for TempWet identified two significantly associated SNPs on chromosome Pv04 and Pv09 (Fig. 8B) tagging 15 genes, of which 14 were located on chromosome Pv09 and one on chromosome Pv04 (Table 3). Among the candidate genes, three of them have putative chloroplast functions, with another gene annotated as a putative peroxidase protein (Table 3). Association analysis for PrecWarm identified two significantly associated SNPs on chromosome Pv01 and Pv08 (Fig. 9A), tagging 14 genes. Interestingly the SNP located on Pv08, and the tagged gene models, were also identified previously as significantly associated with PrecWet. Among the 8 genes located on Pv01 (Table 4), we identified several Arabidopsis homologs, including a putative transcription elongation factor, a pentatricopeptide repeat (PPR) protein, and a receptor-like kinase.

Association analysis for TempWarm identified two significantly associated SNPs, only 25 bp apart, on Pv03 (Fig. 9B) tagging in total 10 gene models with homology to Arabidopsis genes (Table 5). Among those genes, we identified a putative transcriptional activator of small-nuclear RNA genes, a gene possibly involved in cellulose biosynthesis, a chloroplast translation initiation factor, and a gene involved in brassinosteroid signaling.

Environmental and genetic differentiation between Brazilian market classes.

To identify possible donors of drought and heat tolerance, we characterized the distribution of the climatic variables across the different market classes analyzed. An ANOVA analysis showed significant differences across market classes between these climatic variables (Table 6). In particular, the Mulatinho type seems to be cultivated in areas with considerably lower rainfall than the other market classes, with an average rainfall of 485 mm during the wettest quarter and 290 mm during the warmest quarter. A post-hoc HSD test showed a significantly lower PrecWet and PrecWarm for the Mulatinho type (Table 6), even though there was no significant difference for PrecWet with the Preto type (average 533 mm). On the other hand, the Rosinha, Roxo, Pardo, and Manteigão types showed higher rainfalls for both precipitation variables analyzed. Regarding the temperature comparisons for the wettest season (TempWet), the Rosinha type seems to be cultivated at the highest average temperatures (23.3°C), although the Mulatinho, Manteigão and Roxo showed average cultivation temperatures around 23.0°C as well (Table 6). Indeed, an HSD-test showed no significant differences in TempWet for these varieties. The Preto type was cultivated at significantly lower temperatures (average 20°C) than the landraces of the other market classes, except for Pardo and the ‘Others’ market class group. For the temperature comparisons for the warmest season (TempWarm), the Mulatinho type showed the highest average temperature during cultivation period (24.7°C), even though the HSD-test did not show any significant difference between Mulatinho and Manteigão, Rosinha or Roxo for this variable (Table 6).

Table 3

Candidate genes identified by genome-wide association analysis for association with TempWet (BIO8 - Mean temperature of wettest quarter)
GeneID	Arabidopsis homolog	Description
Chromosome Pv04
Phvul.004G106800	ARC11, ATMIND1, MIND	Septum site-determining protein (MIND)
Chromosome Pv09
Phvul.009G085400	WIN2	HOPW1-1-interacting 2
Phvul.009G085500	AT5G24970.2	Protein kinase superfamily protein
Phvul.009G085600	AT5G10745.1
Phvul.009G085700	AT5G10750.1	Enhanced disease resistance-like protein (DUF1336)
Phvul.009G085800	AT2G24800.1	Peroxidase superfamily protein
Phvul.009G085900	ATDBR1, DBR1	Debranching enzyme 1
Phvul.009G086000	TIC55-II	Translocon at the inner envelope membrane of chloroplasts 55-II
Phvul.009G086200	AtMAPR2, MAPR2	Membrane-associated progesterone binding protein 2
Phvul.009G086300	AT2G24960.2	Myb/SANT-like DNA-binding domain protein
Phvul.009G086400	AT4G31860.1	Protein phosphatase 2C family protein
Phvul.009G086500	PGR3	Protein phosphatase 2C family protein
Phvul.009G086600	AT5G25080.1	Sas10/Utp3/C1D family; early nodulin-like protein 14
Phvul.009G086700	AT2G25060.1	AtENODL14, ENODL14
Phvul.009G086800	AT5G10840.1	Endomembrane protein 70 protein family

Table 4

Candidate genes identified by genome-wide association analysis for association with PrecWarm (BIO18 - Precipitation of warmest quarter)
GeneID	Arabidopsis homologous	Description
Chromosome Pv01
Phvul.001G068300	SPT4-2	SPT4 homolog 2
Phvul.001G068400	COD1	Pentatricopeptide repeat (PPR) superfamily protein
Phvul.001G068500	AT1G66910.1	Protein kinase superfamily protein
Phvul.001G068600	LRK10	Receptor-like kinase superfamily protein
Phvul.001G068700
Phvul.001G068800	AT4G18250.1	Receptor serine/threonine kinase, putative
Phvul.001G068900	AT2G01050.1	Zinc ion binding; nucleic acid binding
Phvul.001G069000	AT1G18390.2	Protein kinase superfamily protein
Chromosome Pv08
Phvul.008G166100
Phvul.008G166200	AT5G54970.1
Phvul.008G166300
Phvul.008G166400	AT5G54970.1
Phvul.008G166500	AT1G77700.1	Pathogenesis-related thaumatin superfamily protein
Phvul.008G166600	AT1G38131.1	O-fucosyltransferase family protein

Table 5

Candidate genes identified by genome-wide association analysis for association with TempWarm (BIO10 - Mean temperature of warmest quarter)
GeneID	Arabidopsis homolog	Description
Chromosome Pv03
Phvul.003G154500	SRD2	Small-nuclear RNA activating complex family protein
Phvul.003G154600	CESA7	Cellulose synthase family protein
Phvul.003G154700	SVR9	Chloroplastic translation initiation factor 3 protein
Phvul.003G154800	HSP70B	Heat shock protein 70B
Phvul.003G154900	eIFiso4G1	MIF4G domain-containing protein / MA3 domain-containing protein
Phvul.003G155000	ATPRD2,MPS1,PRD2	Multipolar spindle 1
Phvul.003G155100	SKIP1	SKP1 interacting partner 1
Phvul.003G155200	AT2G24030.1	Zinc ion binding;nucleic acid binding
Phvul.003G155300	AT5G57910.1
Phvul.003G155400	BRS1	Alpha/beta-Hydrolases superfamily protein

Table 6

Differences for temperature and precipitation among geographic distributions of market classes of Brazilian common bean
Market classes (N)	PrecWet (mm)		TempWet (^oC)		PrecWarm (mm)		TempWarm (^oC)
Market classes (N)	Mean	HSD	Mean	HSD	Mean	HSD	Mean	HSD
Manteigão (33)	598	ab	22.9	a	409	a	24.2	ab
Mulatinho (68)	485	c	22.8	a	290	b	24.7	a
Other (32)	588	ab	21.6	ab	460	a	22.9	c
Pardo (23)	621	ab	22.1	ab	496	a	23.2	bc
Preto (70)	533	bc	20.3	b	442	a	22.8	c
Rosinha (14)	615	ab	23.5	a	473	a	23.9	abc
Roxo (17)	692	a	22.9	a	535	a	23.7	abc
PrecWet (BIO16): Precipitation of wettest quarter; TempWet (BIO8): Mean Temperature of wettest quarter; PrecWarm (BIO18): Precipitation of warmest quarter; TempWarm (BIO10): Mean temperature of warmest quarter. HSD: Tukey post hoc test

The GWAS performed to identify signatures of environmental selection could be confounded by cultural preferences for specific seed types, growth habit or other unknown factors related to the different market classes identified in the collection. In order to disentangle true association with environmental factors from farmer or consumer preferences, we performed an F_st-based population differentiation approach using the seven market classes as different populations. This analysis identified several outlier SNPs within these Brazilian market classes, located throughout the Pv11 common bean chromosomes (Fig. 10), with two of them located on chromosomes Pv04 and Pv08 that were identified also by association analysis with TempWet (Pv04), PrecWet and PrecWarm (Pv08).

The recent improvements in genotyping technologies, together with the efforts to sequence and assemble the genomes of several crop plants, have greatly increased the throughput and ability to characterize genetic diversity and population structure in germplasm collections using thousands of genome-wide molecular markers (Metzker 2010; Li et al. 2014; Unamba et al. 2015; Nkhata et al. 2020; Wu et al. 2020; Delfini et al. 2021). This progress could greatly facilitate the pre-screening of large germplasm collections and accelerate the ability to exploit the mostly untapped resources available in public gene banks repositories (Kilian and Graner 2012; Ray and Satya 2014; Ariani et al. 2018). The improvements in sequencing technologies are fostering the development of novel conceptual frameworks for characterizing genetic diversity and understanding environmental adaptation in plant natural populations by integrating the availability of ecological and climatic data with genetic information (Schoville et al. 2012; Bragg et al. 2015; Ariani et al. 2018; Shim et al. 2021).

A Brazilian common bean diversity panel, including 91 landraces and 97 elite lines, was recently characterized using ~ 6,000 genetic markers obtained with DArT-Seq technology by Valdisser et al. (2017), who observed the usual divergence between Andean and Mesoamerican entries (Kwak and Gepts 2009) and a further diversification in the Mesoamerican gene pool among different grain types. More recently, two additional publications describe SNP diversity in specific Brazilian diversity panels of common bean. Delfini et al. (2021), analyzing the genetic diversity of some 50,000 SNPs obtained by GBS (CviAII) among – mostly Mesoamerican − 219 cultivars and breeding lines, observed the usual (Lobaton et al. 2018) strong divergence among Andean and Mesoamerican accessions. Furthermore, they observed a differentiation among groups with different seed coat colors. In a study of a sample of 110 entries consisting predominantly of landraces from more arid, northeastern Brazil, Elias et al. (2021) used GBS (CviAII) and observed that most accessions in the Mesoamerican group clustered according to grain types although hybridizations among clusters within this group were also notable, confirming the results of Burle et al. (2010).

In the current study, we analyzed a larger, more representative collection of Brazilian landraces (274 accessions) using GBS (CviAII)-based SNP markers, which can allow a more precise genetic characterization of this resource compared to SSR markers (Burle et al. 2010, 2011). In contrast, however, with Elias et al. (2021), possible correlations with climatic data were analyzed with bioclimatic variables relevant to the putative growing seasons, i.e., TempWet (BIO8 - Mean temperature of wettest quarter); PrecWet (BIO16 - Precipitation of wettest quarter); TempWarm (BIO10 - Mean temperature of warmest quarter); and PrecWarm (BIO18 - Precipitation of warmest quarter). Although bean growing seasons do not coincide precisely with these quarters, our modified bioclimatic variables provide a more precise approximation of the actual growing seasons. In addition, the integration of genotypic and climatic data for this collection allowed the identification of abiotic stress-resistant market classes and relevant molecular markers for potential improvement by breeding of this crop.

The current analysis of population structure showed a clear separation between the Mesoamerican and Andean gene pools, with great admixture among individuals of the Mesoamerican gene pool. A similar pattern was observed in this same collection previously using microsatellite markers (Burle et al., 2010), but also in other collections of Brazilian germplasm using different marker types (Blair et al., 2013; Valdisser et al., 2017; Delfini et al. 2021). The identification of five sub-populations in the present study agrees with the results of Burle et al. (2011), studying the same bean sample. The absence of admixed individuals between the Mesoamerican and Andean gene pools of common bean could be due to hybrid-weakness observed in F1 crosses between these two gene pools in both wild and domesticated individuals of this species (Gepts and Bliss, 1985; Koinange and Gepts, 1992; Hannah et al., 2007). The F₁ hybrid weakness is the result of two gene-pool-specific loci that are carried mostly by race Mesoamerica, in the Mesoamerican gene pool, and race Nueva Granada, in the Andean pool (Singh et al., 1991). These two eco-geographic races are the main common bean races in Brazil (Burle et al., 2011). The low admixture level between Mesoamerican and Andean genotypes in Brazil could directly affect common bean breeding improvements especially for biotic stress resistance, since it hinders the ability to transfer pathogen resistance genes between the two gene pools, a common strategy used for this purpose (Miklas et al., 2006).

LD decay is a frequently used measure for detecting signatures of selection in specific genomic areas (Barton, 2011). This analysis showed a rapid LD decay for most chromosomes in the Brazilian collection analyzed, possibly facilitating the introgression of novel traits in selected individuals of this collection. However, chromosomes Pv09 and Pv11 showed extensive LD with a half-decay distance of more than 10 Mbp, suggesting a possible selection pressure on these two chromosomes. Indeed, the observed selection sweep could be the result of farmer selection for the most productive plants under standard growing condition in Brazil. One of the major constraints of agricultural production in tropical areas is soil acidity and the ensuing aluminum (Al) toxicity, which highly affects plant productivity by reducing root growth (Panda et al., 2009). Interestingly, more than half of the agricultural areas of Brazil are affected by Al toxicity (Olmos and Camargo, 1976); thus, Brazilian landraces could constitute a useful source of resistance for this stress. The possibility that genomic segments from chromosomes Pv09 and Pv11 could harbor resistance to Al toxicity is supported by a previous study using a bi-parental population of common bean that identified several QTLs affecting root traits in response to Al stress on these chromosomes (López-Marín et al., 2009). An additional possibility is that these chromosomes harbor factors for other traits relevant to Brazilian growing conditions. Elias et al. (2021) identified a major factor for putative drought stress tolerance on Pv09 with, as candidate gene, a Platz transcription factor potentially involved in drought stress tolerance in maize (Zenda et al. 2019).

sPCA is a multivariate test that focuses on the spatial pattern of genetic variability by integrating genetic and geographical data (Jombart et al. 2008). The application of this test can facilitate the identification of long- and short-range genetic patterns in georeferenced plant collection and can allow the correction for population structure in association analysis in the presence of spatial genetic structures (Pyhäjärvi et al. 2013). Previous analysis on wild and domesticated common beans across their native range in the Americas showed a high correlation between geographic and genetic distances, highlighting a significant global structure with physically close genotypes showing high similarities at the genetic level (Papa and Gepts 2003; Rodriguez et al. 2016; Ariani et al. 2018). In contrast, the Brazilian diversity panel analyzed in this study showed extensive local structure with genetically different genotypes sharing similar climatic characteristics (i.e., phenotypes). This phenomenon can reduce the ability to identify variants involved in environmental adaptation, thus drastically reducing the number of significantly associated SNPs. The high local structure observed in the current collection could indeed be one of the reasons why only a few SNPs were significantly associated with environmental data. However, this can also suggest a broad adaptation of the genotypes analyzed, and a high level of hybridization between genetically different varieties. Both results could be beneficial for breeding improvement of local Brazilian landraces and confirm previous results regarding genetic diversity and population structure into this collection (Burle et al. 2010).

Association analysis between genotypic and the environmental data estimated for the cultivation period showed few significantly associated SNPs when correcting for population structure. These results, nevertheless, indicate that signatures of selection from agricultural practices or environmental variations during growing season could be detected using landscape genomic methods also in domesticated common bean collections.

Association analysis with PrecWet identified two SNPs significantly associated with this bio-climatic variable, one located in Pv02 and the other in Pv08, tagging 19 genes in total. Among the gene models identified in Pv02, there are several that could be associated with water stress response based on their annotations (Table 2). As an example, Phvul.002G325900 (annotated as a poly(A) polymerase homologous of Arabidopsis: PAPS1) has been implicated in regulation of plant growth and flowering time in Arabidopsis (Vi et al. 2013; Czesnick and Lenhard 2016). Other interesting genes identified by this analysis include Phvul.002G326000, annotated as a homolog of Arabidopsis CC1, a gene associated with plant growth (mostly through cellulose biosynthesis) and stress resistance (Endler et al. 2015). The Phvul.002G326800 gene model, annotated as a cyclin-dependent kinase, may be related to regulation of cell cycle and plant growth (Umeda et al. 2005). Since the ability to finely regulate flowering time and plant growth is an essential feature for drought resistance or avoidance in this species (Beebe et al. 2013), these genes can constitute interesting candidates for further genetic studies and also for improving these traits in domesticated common bean.

Among the other gene models associated with PrecWet, we identified Phvul.002G326900, a homolog of Arabidopsis gene PMI1, and Phvul.002G326600, an ACC oxidase involved in ethylene biosynthesis (ACC oxidase) and a homolog of Arabidopsis ACO4. The Arabidopsis PMI1 is a gene involved in chloroplast movement in response to environmental signals (especially light-related), a mechanism that is essential for plant adaptive response to stress (Suetsugu et al. 2015) and has also been related to ABA signaling in response to water deficit (Rojas-Pierce et al. 2014). ACC oxidase is the last and rate-limiting enzyme for ethylene biosynthesis (Ruduś et al. 2013), a phytohormone essential in several processes related to plant growth and development, as well as in the response to environmental stresses (Wang et al. 2013; Iqbal et al. 2017). Due to their relevance in important plant physiological processes and stress response, the two genes identified by this analysis could represent interesting molecular markers for improving stress resistance in domesticated common bean.

Association analysis with PrecWarm identified two SNPs associated with this bioclimatic variable, located in Pv01 and Pv08, tagging in total 14 genes. Among the genes identified in Pv01, several of them could be related with responses to environmental variables. As example, the gene model Phvul.001G068300 is a homolog of Arabidopsis SPT4-2, which has been associated with the regulation of auxin response and transcriptional regulation of auxin-related genes in Arabidopsis (Dürr et al. 2014). Auxin is an essential phytohormone involved in multiple developmental and environmental responses (Bielach, Hrtyan, and Tognetti 2017). Other genes associated with PrecWarm identified in this study were Phvul.001G068400, a homolog of the Arabidopsis CYTOCHROME C OXIDASE DEFICIENT 1 (COD1) gene, and Phvul.001G068600, annotated as the Arabidopsis receptor-like kinase LRK10. COD1 is a PPR gene involved in RNA-editing in mitochondrion and is essential for the correct development of the respiratory complex in plants (Dahan et al. 2014), while LRK10 has been related to ABA signaling and drought response in that species (Lim et al. 2015). Due to their involvement in essential cellular processes and drought response, these two genes might also represent interesting candidates for further studies in this species.

Association analysis with TempWet identified two SNPs significantly associated with this bio-climatic variable tagging in total 15 genes, one of them located in Pv04 and the others in Pv09. Based on their annotations, some of the genes identified could be involved in a response to environmental stresses, in particular to temperature and light variations. One of the genes identified (Phvul.009G085800) is annotated as a peroxidase protein. Members of this gene family are involved in several biological processes, mostly as antioxidant enzymes for hydrogen peroxide degradation (Veitch 2004; Pandey et al. 2017). The other genes identified by this association analysis are involved in the regulation of chloroplast development at molecular level. In particular, Phvul.004G106800, a homolog of Arabidopsis ARC11, is involved in chloroplast division (Fujiwara et al. 2008) while the homolog of Phvul.009G086500 (PGR3) is a nuclear gene involved in transcriptional regulation of chloroplast genes (Yamazaki et al. 2004). The other gene model identified by this association analysis (Phvul.009G086000) is homologous to Arabidopsis TIC55-II, a gene involved in chloroplast protein import and chlorophyll degradation in that plant (Boij et al., 2009; Hauenstein et al., 2016). Due to the central role of chloroplasts in response and adaptation to external stimuli in plants, especially light- and water-related signals (Spetea et al. 2014; Dutta et al. 2017; Tamburino et al. 2017), the genes identified by this association analysis would require further investigations for evaluating their role in stress response in common bean.

Association analysis with TempWarm identified two significantly associated SNPs located on Pv03, tagging in total 10 genes. Among these genes, some of them could be related to environmental stress responses, like heat stress. As an example, the gene model Phvul.003G154500, a homolog of Arabidopsis SHOOT REDIFFERENTIATION DEFECTIVE 2 (SRD2), is involved in cell proliferation and in lateral root development in Arabidopsis (Ohtani and Sugiyama 2005; Ohtani, Demura, and Sugiyama 2010); a knockdown mutation of this gene causes aberrant growth phenotypes under high temperatures in this species (Ohtani and Sugiyama 2005). Another gene model identified in this genomic interval is Phvul.003G154600, homolog of Arabidopsis CESA7, a gene involved in cell wall formation in that species (Bosca et al. 2006). Due to the importance of plant cell walls in heat stress response (Wu et al. 2018), this gene could represent an interesting marker for improving heat stress in common bean. Among the additional genes identified in this interval there are Phvul.003G154700, homolog of Arabidopsis SUPPRESSOR OF VARIEGATION 9 (SVR9), and Phvul.003G155400, homolog of Arabidopsis BRI1 SUPPRESSOR 1 (BRS1). SVR9 is essential for chloroplast development and plant survival; it can also affect leaf development in Arabidopsis (Zheng et al. 2016). BRS1 is a suppressor of the Brassinosteroid-Insensitive 1 (BRI1) gene, an essential gene involved in plant growth and development in this species (Li et al. 2001). Due to their involvement in plant growth and development in Arabidopsis, the two common bean homologs identified in this study could represent interesting candidate genes for further genetic studies and validation.

The integration of environmental and genotypic data can facilitate the identification of both useful molecular markers for breeding improvement and suitable donor germplasm of stress resistance that can be used as parents for varietal improvement (Schoville et al. 2012; Henry and Nevo 2014; Berny Mier y Teran et al. 2020). The comparison of climatic variables among Brazilian market classes showed that the landraces of Mulatinho type were cultivated in areas with significantly lower rainfall and higher temperature in Brazil, confirming earlier results obtained with a similar approach using microsatellite markers (Burle 2008). We hypothesize that initial growers of Mulatinho types may have chosen this type because of its relative drought resistance regardless of the seed color. Subsequently, a consumer preference developed for the specific Mulatinho seed color and became the main reason for cultivating this bean type regardless of its drought adaptation. Yet, selection for seed color may have entailed indirect selection for heat or drought tolerance because of linkage disequilibrium, further enhanced by the predominantly autogamous reproductive system of common bean (Papa and Gepts 2003; Papa et al. 2005; Kwak and Gepts 2009). Seed color and color pattern are conditioned by some 7–10 genes distributed over most of the bean chromosomes (Bassett 2007; McClean et al. 2002). Selection for seed color in common bean by farmers is effective as demonstrated by the tremendous biocultural diversity displayed by domesticated common bean (e.g., Kuzay et al. 2020) and shown recently in places as diverse as Mexico (Worthington et al. 2012) and Uganda (Wilkus et al. 2019). The current results confirm the extensive levels of LD in the bean genome identified previously (Kwak and Gepts 2009). These results suggest that some of the SNPs identified by GWAS analysis could be related to environmental adaptation or that there can be linkage between loci differentiating market classes and those controlling mechanisms of environmental adaptation. Nevertheless, caution should be exercised because spurious associations can occur during GWAS analysis (Shen and Carlborg 2013).

In the current study, we performed an extensive analysis of molecular diversity, LD, and population structure in a diversity panel of Brazilian landraces using genome-wide SNP markers obtained with a GBS approach. The integration of geographic (G: climatic data) information, agronomic (A) practice and market class information, combined with molecular (M) genetic data (hence labeled the GAMA approach) allowed us to identify several genes possibly associated with environmental adaptation, as well as possible drought- and heat-tolerant landraces in this collection. This integration would be especially useful in exploiting the genetic resources, both at the accession and genome levels, contained in large germplasm collections provided geographic provenance and molecular diversity data is available.

Plant material, DNA extraction and GBS library preparation

Common bean (P. vulgaris) landraces used in this study were provided by the Embrapa (Empresa Brasileira de Pesquisa Agropecuária) Rice and Bean Unit, located in Santo Antônio de Goiás, GO, Brazil. The collection analyzed in the current study (274 landraces; Table S1) was selected previously, based on passport data, for maximizing the diversity and geographical representation of common bean landraces across bean-growing regions in Brazil (Burle et al. 2010). Earlier analyses provided a genotypic evaluation of SSR diversity (Burle et al. 2010) and an integration of phenotypic and genotypic data (Burle et al. 2011). Genomic DNA (gDNA) was extracted from seeds by pooling the embryos of two dry seeds for each landrace analyzed. The extraction protocol followed the one described in Ariani et al. (2016). Only DNA with an absorbance ratio (A260/A280) > 1.7 (measured with a Nanodrop Lite, Thermo Fisher Scientific) and no sign of degradation on a 1% agarose gel was used for subsequent library preparation. Prior to library preparation, the genomic DNA was quantified with Quant-iT™ PicoGreen® dsDNA Assay Kit (Thermo Fisher Scientific); 100ng of DNA was used for GBS library preparation. The GBS libraries were prepared following the protocol implemented in our laboratory, using CviAII as restriction enzyme (Ariani et al. 2016). The presence of adapter dimers in the sequencing libraries was checked with Experion DNA analysis kit (Biorad, Berkeley, CA, USA). The samples were multiplexed in two libraries with 144 accessions each, including one blank sample for library. The list of the genotypes sequenced in each library, along with passport data, barcode sequences, and SRA run ID, is shown in Table S1.

DNA Sequencing and SNP calling

GBS libraries were sequenced in the DNA Technology Core at the UC Davis Genome Center, using a 100 bp single-end protocol on a HiSeq3000 Illumina sequencing machine. Raw sequencing reads are available under run IDs SRR17267581 and SRR17267582 (see Table S1) in the NCBI short short-read archive database (http://www.ncbi.nlm.nih.gov/sra). Raw reads were preprocessed and demultiplexed with the GBSprep pipeline (https://github.com/aariani/GBSprep) by specifying CviAII recognition and overhang sequences (CATG and ATG, respectively) and by removing overhang restriction sequences from the reads. Reads were then aligned to the P. vulgaris (G19833) v1 reference genome sequence (Schmutz et al. 2014) with BWA (Li and Durbin 2009). After the alignment, only the reads with a minimum mapping quality of 10 were kept for downstream analysis. SNPs were called with NGSEP genotyping framework (Duitama et al. 2014) using only variants with a minimum genotype quality of 40 (-minQuality 40), a maximum number of alignments per starting position of 100 (-maxAlnsPerStartPos 100), and a heterozygosity rate of 0.0001 (-h 0.0001). SNPs were filtered with VCFtools (Danecek et al. 2011) both before and after imputation. Before imputation, we removed the SNPs located in repetitive regions and the genotype calls with a depth lower than 5 (--minDP 5); we kept only sites with at least 10% of the genotype calls passing the first filtering (--max-missing 0.1). For imputation, only SNPs anchored to common bean chromosomes in the reference genome sequence were used. SNPs were imputed with the ImputeVCF plugin of NGSEP specifying a maximum of 8 groups for haplotype clustering (-k 8), and an estimated average recombination rate of 0.004 cM/Kb. This recombination rate was selected based on the average genomic recombination rate estimated from bi-parental populations of common bean (Schmutz et al., 2014). After imputation, we removed genotype calls with a quality less than 10 (--minGQ 10) and with a minor allele frequency less than 0.05 (--maf 0.05). We then removed individuals with more than 50% of missing genotype calls. Finally, only SNPs with at least 80% of the genotype calls passing the final filtering were retained (--max-missing 0.8).

Variant statistics were calculated using VCFtools (Danecek et al., 2011). Nucleotide diversity (π) was calculated on 100kb/20kb sliding windows containing at least 3 variants, and as an average across these windows for each chromosome. Nucleotide diversity (π) was evaluated also on different sub-sets of the genotypes based on the best number of sub-populations estimated by population structure analysis (see below). We performed a separate analysis for ‘pure Mesoamerican’ genotypes (i.e. assigned to one of the Mesoamerican sub-populations), ‘pure Andean’, and ‘Mesoamerican hybrids’ (i.e. hybrids between Mesoamerican sub-populations), using all the SNPs retained after the final filtering step with less than 20% missing data in the different groups.

Genetic principal component analysis (PCA), population structure, and spatial PCA (sPCA)

Genetic PCA was performed using the adegenet package (Jombart 2008) available in the R statistical environment (www.r-project.org). Population structure was evaluated with the sNMF program implemented in the LEA R package (Frichot et al. 2014; Frichot and François 2015). This algorithm was run with the number of sub-populations ranging from 2 to 15, using ten replicated runs for each K, and with the cross-entropy parameter set as True. The population clustering was initiated with two sub-populations (K = 2) because previous analyses had shown the presence of representatives of both the Andean and Mesoamerican gene pools of this domesticated species in Brazil (Gepts et al. 1988; Pereira 1994; Burle et al. 2010). The best number of K was selected based on the number of sub-populations that determine an inflexion point in the cross-entropy profile. Accessions with a membership coefficient < 0.7 were considered hybrids between populations.

Spatial (sPCA) analysis (Jombart et al. 2008) was performed with the spca function available in the adegenet package (Jombart 2008) using a nearest neighbor approach with 10 neighbors. Correlation between genetic and geographic distances of the genotypes´ collecting sites (expressed as geographic coordinates) was evaluated using a randomized Mantel test (Mantel 1967) with 10,000 permutations implemented in the adegenet package (Jombart 2008).

Linkage disequilibrium

Linkage disequilibrium (LD) between SNPs (r²) was estimated with the LDcorSV package (Mangin et al. 2012) of R, a novel method that corrects for bias introduced in r² calculation by population structure and relatedness (kinship) of the sample. We calculated standard r², r² corrected for population structure (r²_s), r² corrected for relatedness (r²_v), and r² corrected for relatedness and population structure (r²_vs). As population structure matrix we used the ancestry coefficient from the sNMF algorithm with 4 sub-populations. The kinship matrix was calculated from the SNP dataset using the EMMA algorithm (Kang et al. 2008) implemented in GAPIT (Lipka et al. 2012). LD distribution across chromosomes was visualized with the LDheatmap package of R (www.r-project.org) using standard r² and corrected values (r²_s, r²_v, and r²_vs). LD decay was estimated using the corrected r² measures (r²_s, r²_v, and r²_vs) with the model proposed by (Hill and Weir 1988), as previously described in Marroni et al. (2011). We performed a similar LD analysis for ‘pure Mesoamerican’ genotypes (i.e., assigned to one of the Mesoamerican sub-populations), ‘pure Andean’, and ‘Mesoamerican hybrids’ (i.e., hybrids among Mesoamerican sub-populations), as identified by population clustering analysis. For each of these groups we evaluated r² corrected by relatedness (r²_v) using SNPs with a MAF higher than 0.01 and with less than 20% of missing data.

GIS analysis

Geographic coordinates for the collecting sites of each genotype analyzed in this study were used for retrieving bioclimatic variables from the WorldClim database (v1.4) (http://www.worldclim.org/) using the dismo package (and its dependencies) within the R statistical environment (www.r-project.org). The bioclimatic variables were downloaded at the maximum resolution available (30 seconds, approximately 0.86 km² at the equator). The bioclimatic variables available in WorldClim represent annual trends (e.g., mean annual temperature, annual precipitation), seasonality (e.g., annual range in temperature and precipitation), and limiting environmental factors (e.g., temperature of the coldest and warmest month, and precipitation of the wettest or driest quarters). To estimate the environmental conditions of common bean growing periods, we used some of the WorldClim bioclimatic variables for quarter periods (three months), instead of annual variables. Common bean is cultivated in Brazil in different periods/months of the year, according to the different production regions and cropping systems (Barbosa and Gonzaga, 2012). However, except for the irrigated system (also called the autumn-winter harvest), which coincides with autumn and/or winter seasons in Brazil, all the other periods of common bean cultivation in the different regions of Brazil coincide with the rainy seasons (no irrigation used) and, at least, part of the summer season. Considering that our study sample is composed exclusively of landraces, grown without irrigation, we used the WorldClim bioclimatic variables related to the wettest quarter and the ones related to the warmest quarter, as follows, respectively: BIO8- Mean temperature of wettest quarter, we named TempWet; BIO16-Precipitation of wettest quarter, named PrecWet; BIO10- Mean temperature of warmest quarter, named TempWarm; and BIO18- Precipitation of warmest quarter, named PrecWarm.

Market class analysis

The collection of 274 Brazilian landraces analyzed in this study were subdivided into common bean market classes (Table S1) by the former Legume Curator Jaime R. Fonseca according to previous classifications (Vieira 1967; Fonseca and Vieira 1986, 2001). To reduce the bias that could be introduced by small groups during statistical and population genetic analysis, the market classes with less than ten landraces were clustered into the ‘Others’ group. ANOVA analysis and Tukey post hoc test (HSD; P = 0.05) were performed to identify variations across market classes for the four above-mentioned bioclimatic variables (PrecWet, TempWet, PrecWarm, and TempWarm) within the R statistical environment (www.r-project.org) using the package agricolae. To identify outlier SNPs that differentiate between market classes and could confound the Genome Wide Association Study (GWAS) analysis, we applied an F_st-based population differentiation approach implemented in Bayescan v2.1 (Foll and Gaggiotti 2008) with the default parameters. A significant FDR threshold of 0.01 was applied to identify significant outlier SNPs. The significant SNPs where then compared with the significant SNPs identified by association analysis.

Association analysis

Association analysis was performed on the climatic data used to estimate the environmental conditions for the cultivation periods of common bean (see previous section) using the FarmCPU algorithm (Liu et al. 2016). Since the hapmap numerical input of FarmCPU does not allow missing data we performed a naive numerical imputation with the Imputer function of the scikit-learn python library (Pedregosa et al. 2011). Briefly, we replaced missing data with the average value (based on numerical hapmap format) of the marker across all the genotypes. FarmCPU is able to detect significant associations and reduce false positive even without the incorporation of population structure or genetic PCA data (Liu et al. 2016). For this reason, we performed the association analysis with and without genetic PCA data, and visually inspected the quantile-quantile (QQ) plot results for identifying the most effective correction method.

Identification of significant SNPs and candidate genes

One of the main problems of GWAS is the high number of association tests that drastically increase the number of false positives SNPs necessitating P-value correction for multiple testing (Shen and Carlborg, 2013). For this reason, we used an approach like the one described by Moghaddam et al. (2016). In brief, we created an empirical distribution of P-values for each GWAS by bootstrapping P-values 10,000 times with replacement; we then considered SNPs to be significant when they fell in the lower 0.01 percentile of the distribution. By using this approach, each association test had different significance cut-offs based on the P-value distribution for each phenotype. Once the significant SNPs were identified for each trait, we selected as candidate genes those within 50Kb from these SNPs (i.e., a 100Kb region centered on the significant SNPs) by using the GenomicRanges/rtracklayer R packages (Lawrence et al., 2009, 2013) and the P. vulgaris (G19833) v1 genome annotation from Phytozome (https://phytozome.jgi.doe.gov/pz/portal.html).

Funding

MLB received a CAPES (Brazil) fellowship. Research funded by USDA/FAS/ICD/RSED and NSF/ERA-CAPS collaborative research No. 1539848.

Competing Interests

The authors declare no competing interests with this research.

Author Contributions

Study concept: PG; plant material: MLB; data collection: AA, MLB; data analysis and interpretation: all authors. First draft: MLB; subsequent versions: all authors. All authors read and approved the final manuscript.

Availability of material and data

Bean accessions are available from the Embrapa Rice and Beans Active Germplasm Bank, located in Santo Antônio de Goiás (germplasm information available at http://alelobag.cenargen.embrapa.br/AleloConsultas/Passaporte/banco.do?idi=20). Sequence data have been deposited at the NCBI short short-read archive database under run IDs SRR17267581 and SRR17267582 (see Table S1) (http://www.ncbi.nlm.nih.gov/sra).

Ethics approval

The plant materials used in this study were obtained from the gene bank of Embrapa CENARGEN under the International Treaty on Plant Genetic Resources for Food and Agriculture.

Abebe TD, Naz AA, Léon J (2015) Landscape genomics reveal signatures of local adaptation in barley (Hordeum vulgare L.). Frontiers in Plant Science 6: 813 https://doi.org/10.3389/fpls.2015.00813
Ariani A, Berny Mier y Teran J, Gepts P (2018) Spatial and temporal scales of range expansion in wild Phaseolus vulgaris. Molecular Biology and Evolution 35: 119–131 https://doi.org/10.1093/molbev/msx273
Ariani A, Berny Mier y Teran JC, Gepts P (2016) Genome-wide identification of SNPs and copy number variation in common bean (Phaseolus vulgaris L.) using genotyping-by-sequencing (GBS). Molecular Breeding 36: 87 https://doi.org/10.1007/s11032-016-0512-9
Barbosa FR, Gonzaga ACdO (2012) Informações técnicas para o cultivo do feijoeiro-comum na Região Central-Brasileira: 2012–2014. In: Ministério da Agricultura PeA (ed). EMBRAPA Arroz e Feijão, Santo Antônio de Goiás, GO, Brasil
Barton NH (2011) Estimating linkage disequilibria. Heredity 106: 205–206 https://doi.org/10.1038/hdy.2010.67
Bassett MJ (2007) Genetics of seed coat color and pattern in common bean. Plant Breeding Reviews 28:239–315. https://doi.org/10.1002/9780470168028.ch8
Beebe SE, Rao IM, Blair MW, Acosta-Gallegos JA (2013) Phenotyping common beans for adaptation to drought. Frontiers in Physiology 4: 35 https://doi.org/10.3389/fphys.2013.00035
Beebe SE, Rao IM, Cajiao C, Grajales M (2008) Selection for drought resistance in common bean also improves yield in phosphorus limited and favorable environments. Crop Science 48:582–592 https://doi.org/10.2135/cropsci2007.07.0404
Berny Mier y Teran JC, Konzen ER, Medina V, Palkovic A, Ariani A, Tsai SM, Gilbert ME, Gepts P (2019a) Root and shoot variation in relation to potential intermittent drought adaptation of Mesoamerican wild common bean (Phaseolus vulgaris L.). Annals of Botany 124:917–932 https://doi.org/10.1093/aob/mcy221
Berny Mier y Teran JC, Konzen ER, Palkovic A, Tsai SM, Rao IM, Beebe S, Gepts P (2019b) Effect of drought stress on the genetic architecture of photosynthate allocation and remobilization in pods of common bean (Phaseolus vulgaris L.), a key species for food security. BMC Plant Biol 19:171 https://doi.org/10.1186/s12870-019-1774-2
Berny Mier y Teran J, Konzen E, Palkovic A, Tsai S, Gepts P (2020) Exploration of the yield potential of Mesoamerican wild common beans from contrasting eco-geographic regions by nested recombinant inbred populations. Frontiers in Plant Science 11:346 https://doi.org/10.3389/fpls.2020.00346
Bielach A, Hrtyan M, Tognetti VB (2017) Plants under stress: Involvement of auxin and cytokinin. International Journal of Molecular Sciences 18: 1427
Bitocchi E, Bellucci E, Giardini A, Rau D, Rodriguez M, Biagetti E, Santilocchi R, Spagnoletti Zeuli P, Gioia T, Logozzo G, Attene G, Nanni L, Papa R (2013) Molecular analysis of the parallel domestication of the common bean (Phaseolus vulgaris) in Mesoamerica and the Andes. New Phytologist 197: 300–313 https://doi.org/10.1111/j.1469-8137.2012.04377.x
Bitocchi E, Nanni L, Bellucci E, Rossi M, Giardini A, Zeuli PS, Logozzo G, Stougaard J, McClean P, Attene G, Papa R (2012) Mesoamerican origin of the common bean (Phaseolus vulgaris L.) is revealed by sequence data. Proceedings of the National Academy of Sciences 109: E788-E796 https://doi.org/10.1073/pnas.1108973109
Blair MW, Brondani RVP, Díaz LM, Del Peloso MJ (2013) Diversity and population structure of common bean from Brazil. Crop Sci 53: 1983–1993 https://doi.org/10.2135/cropsci2012.09.0516
Bosca S, Barton CJ, Taylor NG, Ryden P, Neumetzler L, Pauly M, Roberts K, Seifert GJ (2006) Interactions between MUR10/CesA7-dependent secondary cellulose biosynthesis and primary cell wall structure. Plant Physiology 142: 1353–1363 https://doi.org/10.1104/pp.106.087700
Bragg JG, Supple MA, Andrew RL, Borevitz JO (2015) Genomic variation across landscapes: insights and applications. New Phytologist 207: 953–967 https://doi.org/doi:10.1111/nph.13410
Burle ML (2008) Assessing the genetic diversity of common bean (Phaseolus vulgaris L.) landraces from Brazil: from genetic structure to landscape distribution. PhD Dissertation. University of California, Davis. 342 pp.
Burle ML, Fonseca J, Kami JA, Gepts P (2010) Microsatellite diversity and genetic structure among common bean (Phaseolus vulgaris L.) landraces in Brazil, a secondary center of diversity. Theoretical and Applied Genetics 121: 801–813 https://doi.org/0.1007/s00122-010-1350-5
Burle ML, Fonseca JR, del Peloso MJ, Melo LC, Temple SR, Gepts P (2011) Integrating phenotypic evaluations with a molecular diversity assessment of a Brazilian collection of common bean landraces. Crop Science 51: 2668–2680 https://doi.org/10.2135/cropsci2010.12.0710
Chacón S. MI, Pickersgill B, Debouck DG (2005) Domestication patterns in common bean (Phaseolus vulgaris L.) and the origin of the Mesoamerican and Andean cultivated races. Theoretical and Applied Genetics 110: 432–444 https://doi.org/10.1007/s00122-004-1842-2
Czesnick H, Lenhard M (2016) Antagonistic control of flowering time by functionally specialized poly(A) polymerases in Arabidopsis thaliana. The Plant Journal 88: 570–583 https://doi.org/10.1111/tpj.13280
da Silva OF, Wander AE (2013) O feijão-comum no Brasil: passado, presente e futuro. Documentos, 287. In: Ministério da Agricultura PeA (ed). EMBRAPA Centro Nacional de Pesquisa de Arroz e Feijão, Santo Antônio de Goiás, GO, Brasil, p 63
Dahan J, Tcherkez G, Macherel D, Benamar A, Belcram K, Quadrado M, Arnal N, Mireau H (2014) Disruption of the < em > CYTOCHROME C OXIDASE DEFICIENT1</em > Gene Leads to Cytochrome c Oxidase Depletion and Reorchestrated Respiratory Metabolism in Arabidopsis. Plant Physiology 166: 1788–1802 https://doi.org/10.1104/pp.114.248526
Danecek P, Auton A, Abecasis G, Albers CA, Banks E, DePristo MA, Handsaker RE, Lunter G, Marth GT, Sherry ST, McVean G, Durbin R, Group GPA (2011) The variant call format and VCFtools. Bioinformatics 27: 2156–2158 https://doi.org/10.1093/bioinformatics/btr330
Debouck DG, Toro O, Paredes OM, Johnson WC, Gepts P (1993) Genetic diversity and ecological distribution of Phaseolus vulgaris in northwestern South America. Econ Bot 47: 408–423 https://doi.org/10.1007/s00122-006-0417-9
Delfini J, Moda-Cirino V, Dos Santos Neto J, Ruas PM, Sant’Ana GC, Gepts P, Gonçalves LSA (2021) Population structure, genetic diversity and genomic selection signatures among a Brazilian common bean germplasm. Scientific Reports 11:2694 https://doi.org/10.1038/s41598-021-82437-4
Duitama J, Quintero JC, Cruz DF, Quintero C, Hubmann G, Foulquié-Moreno MR, Verstrepen KJ, Thevelein JM, Tohme J (2014) An integrated framework for discovery and genotyping of genomic variants from high-throughput sequencing experiments. Nucleic Acids Research 42: e44 https://doi.org/10.1093/nar/gkt1381
Dürr J, Lolas IB, Sørensen BB, Schubert V, Houben A, Melzer M, Deutzmann R, Grasser M, Grasser KD (2014) The transcript elongation factor SPT4/SPT5 is involved in auxin-related gene expression in Arabidopsis. Nucleic Acids Research 42: 4332–4347 https://doi.org/10.1093/nar/gku096
Dutta S, Cruz JA, Imran SM, Chen J, Kramer DM, Osteryoung KW (2017) Variations in chloroplast movement and chlorophyll fluorescence among chloroplast division mutants under light stress. Journal of Experimental Botany 68: 3541–3555 https://doi.org/10.1093/jxb/erx203
Elshire RJ, Glaubitz JC, Sun Q, Poland JA, Kawamoto K, Buckler ES, Mitchell SE (2011) A robust, simple genotyping-by-sequencing (GBS) approach for high diversity species. PLoS ONE 6: e19379 https://doi.org/10.1371/journal.pone.0019379
Endler A, Kesten C, Schneider R, Zhang Y, Ivakov A, Froehlich A, Funke N, Persson S (2015) A mechanism for sustained cellulose synthesis during salt stress. Cell 162: 1353–1364 https://doi.org/https://doi.org/10.1016/j.cell.2015.08.028
Foll M, Gaggiotti O (2008) A genome-scan method to identify selected loci appropriate for both dominant and codominant markers: A Bayesian perspective. Genetics 180: 977–993 https://doi.org/10.1534/genetics.108.092221
Fonseca JR, Vieira RF (1986) Algumas características does feijoes plantados nas microrregioes homogeneas 189 e 193 (zona da mata, Minas Gerais). Revista Ceres 33: 449–455 https://doi.org/
Freitas FDO (2006) Genetic-archaeological evidences about the origin of common bean in Brazil. Pesquisa Agropecuaria Brasileira 41: 1199–1203 https://doi.org/10.1590/S0100-204X2006000700018
Freytag GF, Debouck DG (2002) Taxonomy, distribution, and ecology of the genus Phaseolus (Leguminosae — Papilionoideae) in North America, Mexico and Central America. Botanical Research Institute of Texas, Forth Worth, TX
Frichot E, François O (2015) LEA: An R package for landscape and ecological association studies. Methods in Ecology and Evolution 6: 925–929 https://doi.org/doi:10.1111/2041-210X.12382
Frichot E, Mathieu F, Trouillon T, Bouchard G, François O (2014) Fast and efficient estimation of individual ancestry coefficients. Genetics 196: 973–983 https://doi.org/10.1534/genetics.113.160572
Fujiwara MT, Hashimoto H, Kazama Y, Abe T, Yoshida S, Sato N, Itoh RD (2008) The assembly of the FtsZ ring at the mid-chloroplast division site depends on a balance between the activities of AtMinE1 and ARC11/AtMinD1. Plant and Cell Physiology 49: 345–361 https://doi.org/10.1093/pcp/pcn012
Gepts P (1998) Origin and evolution of common bean: past events and recent trends. HortScience 33: 1124–1130 https://doi.org/10.21273/HORTSCI.33.7.1124
Gepts P, Bliss FA (1985) F₁ hybrid weakness in the common bean: differential geographic origin suggests two gene pools in cultivated bean germplasm. J Hered 76: 447–450 https://doi.org/10.1093/oxfordjournals.jhered.a110142
Gepts P, Kmiecik K, Pereira P, Bliss FA (1988) Dissemination pathways of common bean (Phaseolus vulgaris, Fabaceae) deduced from phaseolin electrophoretic variability. I. The Americas. Econ Bot 42: 73–85 https://doi.org/10.1007/bf02859036
Gepts P, Osborn TC, Rashka K, Bliss FA (1986) Phaseolin-protein variability in wild forms and landraces of the common bean (Phaseolus vulgaris): evidence for multiple centers of domestication. Econ Bot 40: 451–468 https://doi.org/10.1007/BF02859659
Hannah MA, Kramer KM, Geffroy V, Kopka J, Blair MW, Erban A, Vallejos CE, Heyer AG, Sanders FET, Millner PA, Pilbeam DJ (2007) Hybrid weakness controlled by the dosage-dependent lethal (DL) gene system in common bean (Phaseolus vulgaris) is caused by a shoot-derived inhibitory signal leading to salicylic acid-associated root death. New Phytologist 176: 537–549 https://doi.org/doi:10.1111/j.1469-8137.2007.02215.x
Henry RJ, Nevo E (2014) Exploring natural selection to guide breeding for agriculture. Plant Biotechnology Journal 12: 655–662 https://doi.org/10.1111/pbi.12215
Hill WG, Weir BS (1988) Variances and covariances of squared linkage disequilibria in finite populations. Theoretical Population Biology 33: 54–78 https://doi.org/https://doi.org/10.1016/0040-5809(88)90004-4
Hoehne FC (1937) Botânica e agricultura no Brasil no seculo XVI: pesquisas e contribuções. Editora Nacional, São Paulo, SP, Brasil
Iqbal N, Khan NA, Ferrante A, Trivellini A, Francini A, Khan MIR (2017) Ethylene role in plant growth, development and senescence: Interaction with other phytohormones. Frontiers in Plant Science 8: 475 https://doi.org/10.3389/fpls.2017.00475
Jombart T (2008) adegenet: a R package for the multivariate analysis of genetic markers. Bioinformatics 24: 1403–1405 https://doi.org/10.1093/bioinformatics/btn129
Jombart T, Devillard S, Dufour AB, Pontier D (2008) Revealing cryptic spatial patterns in genetic variability by a new multivariate method. Heredity 101: 92–103 https://doi.org/10.1038/hdy.2008.34
Kang HM, Zaitlen NA, Wade CM, Kirby A, Heckerman D, Daly MJ, Eskin E (2008) Efficient control of population structure in model organism association mapping. Genetics 178: 1709–1723 https://doi.org/10.1534/genetics.107.080101
Kilian B, Graner A (2012) NGS technologies for analyzing germplasm diversity in genebanks. Briefings in Functional Genomics 11: 38–50 https://doi.org/10.1093/bfgp/elr046
Koenig R, Gepts P (1989) Allozyme diversity in wild Phaseolus vulgaris: further evidence for two major centers of diversity. Theoretical and Applied Genetics 78: 809–817 https://doi.org/10.1007/BF00266663
Koinange EMK, Gepts P (1992) Hybrid weakness in wild Phaseolus vulgaris L. J Hered 83: 135–139 https://doi.org/10.1093/oxfordjournals.jhered.a111173
Kuzay S, Hamilton-Conaty PA, Palkovic A, Gepts P (2020) Is the USDA core collection of common bean representative of genetic diversity of the species, as assessed by SNP diversity? Crop Science 60:1398–1414 https://doi.org/10.2135/cropsci2019.08.0497
Kwak M, Gepts P (2009) Structure of genetic diversity in the two major gene pools of common bean (Phaseolus vulgaris L., Fabaceae). Theoretical and Applied Genetics 118: 979–992 https://doi.org/10.1007/s00122-008-0955-4
Lawrence M, Gentleman R, Carey V (2009) rtracklayer: an R package for interfacing with genome browsers. Bioinformatics 25: 1841–1842 https://doi.org/10.1093/bioinformatics/btp328
Lawrence M, Huber W, Pagès H, Aboyoun P, Carlson M, Gentleman R, Morgan MT, Carey VJ (2013) Software for computing and annotating genomic ranges. PLOS Computational Biology 9: e1003118 https://doi.org/10.1371/journal.pcbi.1003118
Li H, Durbin R (2009) Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics 25: 1754–1760 https://doi.org/10.1093/bioinformatics/btp324
Li J, Lease KA, Tax FE, Walker JC (2001) BRS1, a serine carboxypeptidase, regulates BRI1 signaling in Arabidopsis thaliana. Proceedings of the National Academy of Sciences 98: 5916–5921 https://doi.org/10.1073/pnas.091065998
Li J-Y, Wang J, Zeigler RS (2014) The 3,000 rice genomes project: new opportunities and challenges for future rice research. GigaScience 3:2047–2217X–2043–2048 https://doi.org/10.1186/2047-217X-3-8
Lipka AE, Tian F, Wang Q, Peiffer J, Li M, Bradbury PJ, Gore MA, Buckler ES, Zhang Z (2012) GAPIT: genome association and prediction integrated tool. Bioinformatics 28: 2397–2399 https://doi.org/10.1093/bioinformatics/bts444
Liu X, Huang M, Fan B, Buckler ES, Zhang Z (2016) Iterative usage of fixed and random effect models for powerful and efficient genome-wide association studies. PLOS Genetics 12: e1005767 https://doi.org/10.1371/journal.pgen.1005767
López-Marín HD, Rao IM, Blair MW (2009) Quantitative trait loci for root morphology traits under aluminum stress in common bean (Phaseolus vulgaris L.). Theoretical and Applied Genetics 119: 449–458 https://doi.org/10.1007/s00122-009-1051-0
Mamidi S, Rossi M, Annam D, Moghaddam S, Lee R, Papa R, McClean P (2011) Investigation of the domestication of common bean (Phaseolus vulgaris) using multilocus sequence data. Functional Plant Biology 38: 953–967 https://doi.org/10.1071/fp11124
Mamidi S, Rossi M, Moghaddam SM, Annam D, Lee R, Papa R, McClean PE (2013) Demographic factors shaped diversity in the two gene pools of wild common bean Phaseolus vulgaris L. Heredity 110: 267–276 https://doi.org/10.1038/hdy.2012.82
Mangin B, Siberchicot A, Nicolas S, Doligez A, This P, Cierco-Ayrolles C (2012) Novel measures of linkage disequilibrium that correct the bias due to population structure and relatedness. Heredity 108: 285–291 https://doi.org/10.1038/hdy.2011.73
Mantel N (1967) The detection of disease clustering and a generalized regression approach. Cancer Research 27: 209–220
Marroni F, Pinosio S, Zaina G, Fogolari F, Felice N, Cattonaro F, Morgante M (2011) Nucleotide diversity and linkage disequilibrium in Populus nigra cinnamyl alcohol dehydrogenase (CAD4) gene. Tree Genetics & Genomes 7: 1011–1023 https://doi.org/10.1007/s11295-011-0391-5
McClean PE, Lee RK, Otto C, Gepts P, Bassett MJ (2002) Molecular and phenotypic mapping of genes controlling seed coat pattern and color in common bean (Phaseolus vulgaris L.). Journal of Heredity 93:148–152. https://doi.org/10.1093/jhered/93.2.148
Melo LC (2009) Procedimentos para condução de experimentos de Valor de Cultivo e Uso em feijoeiro comum, Documentos 239. In: Ministério da Agricultura, Pecuária e Abastecimento (ed). EMBRAPA Arroz e Feijão, Santo Antônio de Goiás, GO, Brasil
Metzker ML (2010) Sequencing technologies - the next generation. Nat Rev Genet 11: 31–46 https://doi.org/10.1038/nrg2626
Miklas PN, Kelly JD, Beebe SE, Blair MW (2006) Common bean breeding for resistance against biotic and abiotic stresses: from classical to MAS breeding. Euphytica 147: 106–131 https://doi.org/
Moghaddam SM, Mamidi S, Osorno JM, Lee R, Brick M, Kelly J, Miklas P, Urrea C, Song Q, Cregan P, Grimwood J, Schmutz J, McClean PE (2016) Genome-wide association study identifies candidate loci underlying agronomic traits in a Middle American diversity panel of common bean. The Plant Genome 9:3 https://doi.org/10.3835/plantgenome2016.02.0012
Nkhata W, Shimelis H, Melis R, Chirwa R, Mzengeza T, Mathew I, Shayanowako A (2020) Population structure and genetic diversity analyses of common bean germplasm collections of East and Southern Africa using morphological traits and high-density SNP markers. PLOS ONE 15:e0243238 https://doi.org/10.1371/journal.pone.0243238
Ohtani M, Demura T, Sugiyama M (2010) Particular significance of SRD2-dependent snRNA accumulation in polarized pattern generation during lateral root development of Arabidopsis. Plant and Cell Physiology 51: 2002–2012 https://doi.org/10.1093/pcp/pcq159
Ohtani M, Sugiyama M (2005) Involvement of SRD2-mediated activation of snRNA transcription in the control of cell proliferation competence in Arabidopsis. The Plant Journal 43: 479–490 https://doi.org/10.1111/j.1365-313X.2005.02469.x
Olmos ILJ, Camargo MN (1976) Ocorrência de alumínio tóxico nos solos do Brasil, sua caracterização e distribuição [Incidence of aluminium toxicity in Brazilian soils, its characterization and distribution]. Ciencia e Cultura 28: 171–180
Panda SK, Baluska F, Matsumoto H (2009) Aluminum stress signaling in plants. Plant Signal Behav 4: 592–597 https://doi.org/10.4161/psb.4.7.8903
Pandey VP, Awasthi M, Singh S, Tiwari S, Dwivedi UN (2017) A comprehensive review on function and application of plant peroxidases. Biochemistry & Analytical Biochemistry 6: 308 https://doi.org/10.4172/2161-1009.1000308
Papa R, Gepts P (2003) Asymmetry of gene flow and differential geographical structure of molecular diversity in wild and domesticated common bean (Phaseolus vulgaris L.) from Mesoamerica. Theoretical and Applied Genetics 106: 239–250 https://doi.org/10.1007/s00122-002-1085-z
Parker TA, Berny Mier y Teran JC, Palkovic A, Jernstedt J, Gepts P (2020) Pod indehiscence is a domestication and aridity resilience trait in common bean. New Phytologist 225:558–570 https://doi.org/10.1111/nph.16164
Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V, Vanderplas J, Passos A, Cournapeau D, Brucher M, Perrot M, Duchesnay É (2011) Scikit-learn: machine learning in Python. Journal of Machine Learning Research 12: 2825–2830
Pereira PAA (1990) Evidências de domesticação e disseminação do feijoeiro comum e conseqüências para o melhoramento genético da espécie. Pesq Agropec Bras 25: 19–23
Petry N, Boy E, Wirth JP, Hurrell RF (2015) The potential of the common bean (Phaseolus vulgaris) as a vehicle for iron biofortification. Nutrients 7: 1144–1173 https://doi.org/10.3390/nu7021144
Pyhäjärvi T, Hufford MB, Mezmouk S, Ross-Ibarra J (2013) Complex patterns of local adaptation in teosinte. Genome Biology and Evolution 5: 1594–1609 https://doi.org/10.1093/gbe/evt109
Ray S, Satya P (2014) Next generation sequencing technologies for next generation plant breeding. Frontiers in Plant Science 5: https://doi.org/10.3389/fpls.2014.00367
Rendón-Anaya M, Montero-Vargas JM, Saburido-Alvarez S, Vlasova A, Capella-Gutiérrez S, Ordaz-Ortiz JJ, Aguilar OM, Vianello-Brondani RP, Santalla M, Delaye L, Gabaldón T, Gepts P, Winkler R, Guigó R, Delgado-Salinas A, Herrera-Estrella A (2017) Genomic history of the origin and domestications of common bean in the Americas unveils its closest sister species. Genome Biology 18: 60 https://doi.org/10.1186/s13059-017-1190-6
Rodriguez M, Rau D, Bitocchi E, Bellucci E, Biagetti E, Carboni A, Gepts P, Nanni L, Papa R, Attene G (2016) Landscape genetics, adaptive diversity and population structure in Phaseolus vulgaris. New Phytologist 209: 1781–1794 https://doi.org/doi:10.1111/nph.13713
Rojas-Pierce M, Whippo CW, Davis PA, Hangarter RP, Springer PS (2014) PLASTID MOVEMENT IMPAIRED1 mediates ABA sensitivity during germination and implicates ABA in light-mediated chloroplast movements. Plant Physiology and Biochemistry 83: 185–193 https://doi.org/https://doi.org/10.1016/j.plaphy.2014.07.014
Ruduś I, Sasiak M, Kępczyński J (2013) Regulation of ethylene biosynthesis at the level of 1-aminocyclopropane-1-carboxylate oxidase (ACO) gene. Acta Physiologiae Plantarum 35: 295–307 https://doi.org/10.1007/s11738-012-1096-6
Schmutz J, McClean P, Mamidi S, Wu G, Cannon S, Grimwood J, Jenkins J, Shu S, Song Q, Chavarro C, Torres-Torres M, Geffroy V, Moghaddam SM, Gao D, Abernathy B, Barry K, Blair M, Brick MA, Chovatia M, Gepts P, Goodstein DM, Gonzales M, Hellsten U, Hyten DL, Jia G, Kelly JD, Kudrna D, Lee R, Richard MMS, Miklas PN, Osorno JM, Rodrigues J, Thareau V, Urrea CA, Wan M, Yu Y, Zhang M, Wing RA, Cregan PB, Rokhsar DS, Jackson SA (2014) A reference genome for common bean and genome-wide analysis of dual domestications. Nature Genetics 46: 707–713 https://doi.org/10.1038/ng.3008
Schoville SD, Bonin A, François O, Lobreaux S, Melodelima C, Manel S (2012) Adaptive genetic variation on the landscape: methods and cases. Annual Review of Ecology, Evolution, and Systematics 43: 23–43 https://doi.org/doi:10.1146/annurev-ecolsys-110411-160248
Shen X, Carlborg Ö (2013) Beware of risk for increased false positive rates in genome-wide association studies for phenotypic variability. Frontiers in Genetics 4: https://doi.org/10.3389/fgene.2013.00093
Singh SP, Gepts P, Debouck DG (1991) Races of common bean (Phaseolus vulgaris L., Fabaceae). Econ Bot 45: 379–396 https://doi.org/10.1007/BF02887079
Spetea C, Rintamäki E, Schoefs B (2014) Changing the light environment: chloroplast signalling and response mechanisms. Philosophical Transactions of the Royal Society B: Biological Sciences 369: 20130220 https://doi.org/doi:10.1098/rstb.2013.0220
Suetsugu N, Higa T, Kong S-G, Wada M (2015) PLASTID MOVEMENT IMPAIRED1 and PLASTID MOVEMENT IMPAIRED1-RELATED1 Mediate Photorelocation Movements of Both Chloroplasts and Nuclei. Plant Physiology 169: 1155–1167 https://doi.org/10.1104/pp.15.00214
Tamburino R, Vitale M, Ruggiero A, Sassi M, Sannino L, Arena S, Costa A, Batelli G, Zambrano N, Scaloni A, Grillo S, Scotti N (2017) Chloroplast proteome response to drought stress and recovery in tomato (Solanum lycopersicum L.). BMC Plant Biology 17: 40 https://doi.org/10.1186/s12870-017-0971-0
Umeda M, Shimotohno A, Yamaguchi M (2005) Control of cell division and transcription by cyclin-dependent kinase-activating kinases in plants. Plant and Cell Physiology 46: 1437–1442 https://doi.org/10.1093/pcp/pci170
Unamba CIN, Nag A, Sharma RK (2015) Next generation sequencing technologies: The doorway to the unexplored genomics of non-model plants. Frontiers in Plant Science 6: https://doi.org/10.3389/fpls.2015.01074
Valdisser PAMR, Pereira WJ, Almeida Filho JE, Müller BSF, Coelho GRC, de Menezes IPP, Vianna JPG, Zucchi MI, Lanna AC, Coelho ASG, de Oliveira JP, Moraes AdC, Brondani C, Vianello RP (2017) In-depth genome characterization of a Brazilian common bean core collection using DArTseq high-density SNP genotyping. BMC Genomics 18: 423 https://doi.org/10.1186/s12864-017-3805-4
Veitch NC (2004) Structural determinants of plant peroxidase function. Phytochemistry Reviews 3: 3–18 https://doi.org/10.1023/B:PHYT.0000047799.17604.94
Vi SL, Trost G, Lange P, Czesnick H, Rao N, Lieber D, Laux T, Gray WM, Manley JL, Groth D, Kappel C, Lenhard M (2013) Target specificity among canonical nuclear poly(A) polymerases in plants modulates organ growth and pathogen response. Proceedings of the National Academy of Sciences 110: 13994–13999 https://doi.org/10.1073/pnas.1303967110
Vieira C (1967) O feijoeiro-comum: cultura, doenças e melhoramento. Universidade Federal de Viçosa, Viçosa, MG, Brasil
Vieira C (1988) Phaseolus genetic resources and breeding in Brazil. In: Gepts P (ed) Genetic resources of Phaseolus beans. Kluwer, Dordrecht, the Netherlands, pp 467–483
Vieira C, de Paula Júnior TJ, Borém A (1998) Feijão: Aspectos gerais e cultura no Estado de Minas. Imprensa Universitária, Viçosa, MG, Brasil
Vlasova A, Capella-Gutiérrez S, Rendón-Anaya M, Hernández-Oñate M, Minoche AE, Erb I, Câmara F, Prieto-Barja P, Corvelo A, Sanseverino W, Westergaard G, Dohm JC, Pappas GJ, Saburido-Alvarez S, Kedra D, Gonzalez I, Cozzuto L, Gómez-Garrido J, Aguilar-Morón MA, Andreu N, Aguilar OM, Garcia-Mas J, Zehnsdorf M, Vázquez MP, Delgado-Salinas A, Delaye L, Lowy E, Mentaberry A, Vianello-Brondani RP, García JL, Alioto T, Sánchez F, Himmelbauer H, Santalla M, Notredame C, Gabaldón T, Herrera-Estrella A, Guigó R (2016) Genome and transcriptome analysis of the Mesoamerican common bean and the role of gene duplications in establishing tissue and temporal specialization of genes. Genome Biology 17: 1–18 https://doi.org/10.1186/s13059-016-0883-6
Wang F, Cui X, Sun Y, Dong C-H (2013) Ethylene signaling and regulation in plant growth and stress responses. Plant Cell Reports 32: 1099–1109 https://doi.org/10.1007/s00299-013-1421-6
Westengen OT, Berg PR, Kent MP, Brysting AK (2012) Spatial structure and climatic adaptation in African maize revealed by surveying SNP diversity in relation to global breeding and landrace panels. PLOS ONE 7: e47832 https://doi.org/10.1371/journal.pone.0047832
Wilkus EL, Berny Mier y Teran JC, Mukankusi CM, Gepts P (2018) Genetic patterns of common-bean seed acquisition and early-stage adoption among farmer groups in western Uganda. Frontiers in Plant Science 9:586 https://doi.org/10.3389/fpls.2018.00586
Worthington M, Soleri D, Aragón-Cuevas F, Gepts P (2012) Genetic composition and spatial distribution of farmer-managed bean plantings: an example from a village in Oaxaca, Mexico. Crop Science 52:1721–1735 https://doi.org/10.2135/cropsci2011.09.0518
Wu H-C, Bulgakov VP, Jinn T-L (2018) Pectin methylesterases: Cell wall remodeling proteins are required for plant response to heat stress. Frontiers in plant science 9: 1612–1612 https://doi.org/10.3389/fpls.2018.01612
Yamazaki H, Tasaka M, Shikanai T (2004) PPR motifs of the nucleus-encoded factor, PGR3, function in the selective and distinct steps of chloroplast gene expression in Arabidopsis. The Plant Journal 38: 152–163 https://doi.org/10.1111/j.1365-313X.2004.02035.x
Zenda T, Liu S, Wang X, Liu G, Jin H, Dong A, Yang Y, Duan H (2019) Key maize drought-responsive genes and pathways revealed by comparative transcriptome and physiological analyses of contrasting inbred lines. International Journal of Molecular Sciences 20:1268 https://doi.org/10.3390/ijms20061268
Zheng M, Liu X, Liang S, Fu S, Qi Y, Zhao J, Shao J, An L, Yu F (2016) Chloroplast translation initiation factors regulate leaf variegation and development. Plant Physiology 172: 1117–1130 https://doi.org/10.1104/pp.15.02040

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

The GAMA Approach for the Analysis of Large Germplasm Collections: Genetic Diversity and Landscape Genomics of Common Bean (Phaseolus Vulgaris) Landraces in Brazil

Status:

Version 1

Abstract

Figures

Introduction

Results

Discussion

Conclusion

Materials And Methods

Plant material, DNA extraction and GBS library preparation

DNA Sequencing and SNP calling

Genetic principal component analysis (PCA), population structure, and spatial PCA (sPCA)

Linkage disequilibrium

GIS analysis

Market class analysis

Association analysis

Identification of significant SNPs and candidate genes

Declarations

Funding

Competing Interests

Author Contributions

Availability of material and data

Ethics approval

References

Additional Declarations

Supplementary Files

Status:

Version 1