Towards the Development, Maintenance, and Standardized Phenotypic Characterization of Single‐Seed‐Descent Genetic Resources for Common Bean

The optimal use of legume genetic resources represents a key prerequisite for coping with current agriculture‐related societal challenges, including conservation of agrobiodiversity, agricultural sustainability, food security, and human health. Among legumes, the common bean (Phaseolus vulgaris) is the most economically important for human consumption, and its evolutionary trajectories as a species have been crucial to determining the structure and level of its present and available genetic diversity. Genomic advances are considerably enhancing the characterization and assessment of important genetic variants. For this purpose, the development and availability of, and access to, well‐described and efficiently managed genetic resource collections that comprise pure lines derived by single‐seed‐descent cycles will be paramount for the use of the reservoir of common bean variability and for the advanced breeding of legume crops. This is one of the main aims of the new and challenging European project INCREASE, which is the implementation of Intelligent Collections with appropriate standardized protocols that must be characterized, maintained, and made available, along with the related data, to users such as breeders and researchers. © 2021 The Authors. Current Protocols published by Wiley Periodicals LLC.

Thus, as the first step, PGRs and their wide diversity need to be maintained. Germplasm banks can guarantee the conservation ex situ of such biodiversity. Genesys PGR is a free online global portal that allows the exploration of plant species diversity through a single website (accessible at www.genesys-pgr.org). The data published on Genesys follow the standards for Multi-Crop Passport Descriptors (as defined by the United Nations Food and Agriculture Organization [FAO] and Bioversity). For P. vulgaris, Genesys contains over 135,500 accessions (as of October 19, 2020), which are stored in several holding institutions worldwide. Most of these accessions are landraces (∼71,000), followed by improved cultivars (∼20,000), and wild (∼2000) and breeding/research materials (∼2600). Figure 2 shows the geographic distribution of the wild and landrace accessions of common bean for which information on geographic coordinates is available in Genesys. The International Centre for Tropical Agriculture (CIAT) in Colombia holds the largest P. vulgaris collection, with ∼32,000 accessions, followed by the United States Department of Agriculture Agricultural Research Service (USDA-ARS), with nearly 14,000 accessions, and the Leibniz Institute of Plant Genetics and Crop Plant Research (IPK) in Germany, with ∼8400 accessions. Numerous common bean accessions (∼27,700 overall) are also held by a Brazilian gene bank (the Empresa Brasileira de Pesquisa Agropecuária, or EM-BRAPA). Recursos Genéticos e Biotecnologia CENARGEN, Brasília, DF, and Arroz e Feijão, CNPAF, Goiânia, GO). Furthermore, the Russian collection of common beans at the N.I. Vavilov Institute of Plant Genetic Resources (VIR) is one of the oldest in the world. The VIR collection now numbers ∼6400 accessions (according to information in Genesys) and is the result of the collective efforts of several generations of scientists and explorers, beginning in the early twentieth century.
P. vulgaris germplasm collections stored in gene banks all around the world are essential for national and international attempts to ensure the safety (i.e., conservation and maintenance) and use of such PGRs.

Genomics of Genetic Resources
The genomics era has led to a rapid increase in available sequence data, which has thus provided a more detailed picture of the genetic diversity and structure of crop germplasm, as well as enabling the identification of genetic variants that are the bases of important heritable target traits (Luikart et al., 2018). Genome-wide fingerprinting of single-seed descent (SSD) gene-bank accessions allowed the characterization of a complete barley collection (Milner et al., 2019) and resulted in further identification of potential duplicated material within (Milner et al., 2019) and across (Singh et al., 2019) collections. Moreover, the accurate description of the genetic topology of a gene pool, as revealed by principal-component analysis or other population genomics approaches, allows the identification of hitherto under-represented parts of the overall genetic diversity, and can suggest strategies for completing a collection. Furthermore, once a collection is fingerprinted, existing historic phenotype data obtained during seed multiplication can be used. In the case of barley, the statistical analysis of historic data of key agronomic traits resulted in high heritability estimates for the collection (González et al., 2018), so that users can make informed selections of lines of interest.
For the common bean, the current availability of high-throughput sequencing platforms has allowed the release of high-quality reference genomes of the Andean G19833 (Schmutz et al., 2014) and Mesoamerican BAT93 (Vlasova et al., 2016) genotypes. A further high-quality common bean reference genome of the race Durango pinto UI111 genotype was also released recently (P. vulgaris UI111 v1.1, DOE-JGI and USDA-NIFA, http:// phytozome.jgi.doe.gov/ ). Genetic characterization through high-throughput sequencing of common bean collections and segregant populations is increasing in the literature. One major study that aims to characterize a very wide sample of common bean genetic resources is the BEAN_ADAPT project, which is funded through the second Cortinovis et al.

of 28
Current Protocols European Research Area Network for Coordinating Action in Plant Sciences (ERA-CAPS) joint funding call. As PGRs are often heterogeneous, working with them has required the development of SSD seeds to obtain purified and genetically homogeneous material that is suitable for genetic studies. Therefore, BEAN_ADAPT is based on the characterization of nested core collections, developed by SSD, that have been genotyped at different levels on the basis of their size. The Pv_ALL collection, a large collection of ∼10,000 accessions (mostly European and from the American centers of origin), and a subsample of 500 lines (Pv_core1) have both been characterized through genotyping-bysequencing methods, while a further subsample of 220 accessions have been genotyped by whole-genome sequencing. The genetic data are now being used in combination with the different phenotypic evaluations to identify the genetic basis of phenotypic variants related to adaptation to new environments, through the study of the introduction and wide spread of the common bean from the Americas into Europe. Moreover, the overall design of the project and the data obtained can be also used to predict the phenotypes of the whole collection based on the genotype data, resulting in a very useful tool for exploiting the huge number of PGRs that are currently maintained in gene banks and that remain widely underexploited.
Several genomic studies have been performed based on two particular panels: the Middle American Diversity Panel, which was developed as part of the USDA-funded Common Bean Coordinated Agricultural Project (BeanCAP) and consists of 280 modern bean genotypes that are mostly from the race Mesoamerica, but also from Durango and Jalisco (Moghaddam et al., 2016); and the Andean Diversity Panel, which consists of 396 accessions, the majority of which belong to the Andean gene pool (349), while 21 are Mesoamerican, and 26 are from admixtures between the Mesoamerican and Andean gene pools (Chichy, Porch, et al., 2015). Moghaddam et al. (2016) genotyped the Middle American Diversity Panel with two Illumina iSelect 6K gene chip sets (BAR-CBEAN6K_1 and BARCBEAN6K_2; Song et al., 2015) and two low-pass sequencing protocols (Elshire et al., 2011;Schröder et al., 2016). Overall, the panel was genotyped at 217,486 mapped single-nucleotide polymorphisms (SNPs). These data were combined with phenotype evaluations for classical agronomic traits (carried out in four different American locations), which were used to perform genome-wide association studies (GWAS). This identified new and known genomic regions that affect various agronomic traits, such as days to flowering, days to maturity, growth habit, lodging, canopy height, and seed weight (Moghaddam et al., 2016).
The Andean Diversity Panel was developed and genotyped at 5398 SNPs using the Illumina BARCBean6K_3 SNP BeadChip (Cichy, Porch, et al., 2015). The SNP data were used to genetically characterize the panel by investigating the level and structure of its genetic diversity, and they were also coupled with phenotypic data (i.e., plant determinacy) to identify marker-trait associations. These panels (or subsamples of them, or as part of larger collections) have been used in further GWAS studies (Cichy, Wiesinger, & Mendoza, 2015;Kamfwa, Cichy, & Kelly, 2015a, 2015bMcClean et al., 2017McClean et al., , 2018Moghaddam et al., 2017;Oladzad, Zitnick-Anderson, et al., 2019;Parker et al., 2020;Soltani et al., 2017;Tock et al., 2017;Zuiderveen, Padder, Kamfwa, Song, & Kelly, 2016), demonstrating the potential of such materials for pre-breeding and breeding studies. Recently, Almeida et al. (2020) also efficiently used SNPs to characterize the diversity of an association panel of Carioca strains from Brazil. groups with limited resources; however, at the same time, they are large enough to define the genetic variance at the basis of phenotypic variability. Oladzad, Porch, et al. (2019) carried out genotyping by sequencing for the Bean Abiotic Stress Evaluation panels, and they analyzed all of the data to identify markers associated with production traits under both heat-and drought-stress environments.
Collections and/or panels have also been developed with a focus on snap bean genetic resources. Within the BeanCAP project, a snap bean diversity panel was developed that consisted of 149 cultivars from North America and Europe (Kleintop, Myers, Echeverria, Thompson, & Brick, 2016). Wallace, Arkwazee, Vining, and Myers (2018) genotyped the BeanCAP Snap Bean Panel along with a further 55 Chinese and 19 Spanish snap bean genotypes and 24 heirloom beans at 5398 SNPs. This study provided a deep investigation of the genetic diversity and structure of this collection, which represents a very useful tool for further GWAS. Myers et al. (2019) characterized this panel for total phenolic contents and genotyped them at 10,073 SNPs, and they used GWAS for identification of 11 quantitative trait nucleotides associated with this trait.
Another interesting example of the use of whole-genome scan analysis to characterize P. vulgaris genetic resources was conducted by Wu et al. (2020) and focused on the variability of Chinese common bean germplasm compared to worldwide accessions, and on the relevant contributions of landrace material. They applied whole-genome resequencing to a collection of 683 accessions, which comprised 529 landraces and 154 breeding lines that were representative of both the Mesoamerican and Andean gene pools. The plant material was grown over 3 years and in four locations in China that are characterized by different agro-ecological conditions, with the aim of investigating trait associations related to P. vulgaris yield across the north-south geographic clines. Wu et al. (2020) detected a total of 4,811,097 SNPs and several genetic differences in terms of structural variations affecting DNA segments of >1 kbp, such as insertions, deletions, copy number variations, and presence/absence variations. By using GWAS, these researchers defined several marker-trait associations that were dispersed across the entire genome for all of the traits considered, even though they were found in different proportions depending on the inheritance level and the complexity of the trait itself.
Segregant populations also represent valuable PGRs. As an example of the use of wild individuals to introgress new variability into elite germplasm, for instance for the common bean, Murgia et al. (2017) phenotypically characterized a very interesting introgression line (IL) population. This population was developed from an initial cross between the domesticated Andean variety Midas and the G12873 line, a wild Mesoamerican accession (Koinage et al., 1996). Several cycles of selfing and subsequent backcrosses with the recurrent parent Midas allowed the development of a set of 70 ILs from BC 3 /F 4 :F 5 families and 217 ILs from BC 3 /F 6 :F 7 families Rau et al., 2019). Phenotypic characterization for pod shattering , combined with genetic characterization through genotyping by sequencing at 14,196 SNPs, allowed Rau et al. (2019) to investigate the genetic architecture of the shattering trait in common bean. They identified a locus in the distal part of chromosome Pv05 as the primary locus responsible for the pod indehiscence phenotype, along with numerous other secondary quantitative trait loci that contributed to the modulation of this phenotype. Recently, Di  continued the development of this IL population: from the Murgia et al. (2017) IL population, they selected six highly shattering ILs and used these as donor parents for high pod shattering for further backcrosses (BC 4 ) with Midas, to provide an IL population of 1197 BC 4 /F 4 individuals that was then genetically (19,420 SNPs) and phenotypically evaluated for pod shattering. Along with gene expression and parallel histological analysis of dehiscent and indehiscent pods, a GWAS by Di  identified an ortholog of AtMYB26 from Arabidopsis thaliana as the best candidate for loss of Cortinovis et al.

of 28
Current Protocols pod shattering, in a genomic region ∼11 kb downstream of the most highly associated peak.

Development and Maintenance of Common Bean Increase Intelligent Collections
Several drawbacks characterize conventional conservation management of PGRs. Among these, there is the problem that seed collections are assembled and maintained on an accession basis, with each accession usually constituting a mixture of genotypes that represents a population. As a consequence, the information collected at the phenotype level cannot be directly linked to a specific genotype. Furthermore, hundreds of accessions are conserved and maintained in gene banks with very little information available (i.e., there is a lack of comprehensive information regarding passport data and descriptors useful for users, combined with accession heterogeneity and unharmonized data), which makes their selection and use for specific purposes by researchers and breeders often difficult. Moreover, the available information is not easily accessible, being in databases that are centralized and were not designed to integrate data obtained by external users.
Recently, European Union Horizon 2020 funded Project INCREASE-Intelligent Collections of Food Legumes Genetic Resources for European Agrofood Systems-with the aim of overcoming such limitations and implementing a new approach for conservation, management, and characterization of PGRs (Bellucci et al., submitted).
Along with chickpea, lentil, and lupin (Kroc et al., in preparation; Guerra-García, Gioia, von Wettberg, Logozzo, & Bett, in preparation; Kumar et al., in preparation), common bean is one of the legume species that form the basis of the INCREASE project. With the purpose of conserving, managing, and making the best use of food-legume genetic resources, the INCREASE project is developing innovative conservation procedures through the integration of two complementary strategies (i.e., ex situ and in situ approaches) and the development of SSD-purified accessions based on single homozygous genotypes. SSD collections provide the possibility of associating phenotypes with reliable genotype information, which can then be used to advance and improve legume research, breeding, and cultivation.
To achieve this goal, the project plans to develop INCREASE Intelligent Collections ("intelligent" in terms of being able to memorize, learn, improve, and evolve) as a set of three nested core collections that will be characterized genetically and/or phenotypically at different levels, according to their size. These comprise: • The Reference Core (R-CORE), as the largest collection. For common bean, this will rely on already available SSD lines from previous projects (e.g., BEAN_ADAPT, BRESOV) complemented with heterogeneous accessions conserved in situ and ex situ, from which SSD lines will be developed to provide a worldwide and highly diverse panel of wild accessions, landraces, and cultivars. This collection will include more than 10,000 SSD lines that will be genotyped using at least a low-coverage approach (e.g., genotyping by sequencing and exome capture), and will be maintained in gene banks for long-term conservation. • The Training Core (T-CORE), representing a subsample of the R-CORE and comprising ∼450 lines. A deeper sequencing approach is planned for T-CORE (e.g., Illumina whole-genome sequencing), along with broad phenotypic characterization (e.g., classical, molecular, high-throughput phenotyping) under both controlled and field conditions. • The Hyper-CORE (H-CORE), which will consist of 40 to 80 accessions that are carefully chosen on an evolutionary transect, with the primary aim being to sample the largest possible genotypic diversity of P. vulgaris, including closely related Cortinovis et al.

of 28
Current Protocols species. These lines will be deeply phenotyped and genotyped, and a subsample of the H-CORE will be also sequenced to develop a pangenome.
The large amount of data produced will be analyzed to (i) investigate the level and structure of the genetic diversity of legume genetic resources; (ii) identify functional variants that might have major roles in determining the phenotypic variation for a large number of traits; and (iii) predict the phenotypes of PGRs in gene banks only on the basis of their genotypic characterization. This will provide an incredible tool in the hands of geneticists and breeders for the improvement of food legumes.
Here, we describe the procedures that will be applied in INCREASE to develop the Intelligent Collections, with a specific focus on the protocols that INCREASE has implemented and proposed for adoption by gene banks and research institutions to obtain SSD lines, to conserve and maintain their seeds, to make the seeds available for users, and to characterize them and integrate the data obtained into a centralized system, which will ultimately be accessible to anyone interested in these PGRs. Figure 3 shows the workflow established by INCREASE in terms of the SSD development, conservation, maintenance, and characterization procedures. The protocols developed within INCREASE refer to activities related to Work Package 3 of the project, entitled "Sampling core collections, SSD development, DNA extraction and seed distribution." In particular, the three different phenotyping protocols described below were established to characterize the genetic resources of the project during SSD development and the subsequent seed increase cycles that will be performed under controlled conditions. These protocols were developed starting from IPGRI P. vulgaris descriptors (International Board for Plant Genetic Resources [IBPGR], 1982) and Crop Ontology (Bioversity International, 2011), and they have been modified specifically for purposes related to these SSD development activities.

CHARACTERIZING COMMON BEAN SEEDS FOR SEED TRAIT DESCRIPTORS
This protocol was developed to characterize seeds before the start of any cycle of seed increase. It allows for the characterization of the original phenotypes of seeds from heterogeneous materials, if the cycle to be carried out is the first (development of SSD lines from heterogeneous accessions), and for obtaining data on the seed morphology of each SSD line for each subsequent cycle of seed multiplication. Such data are important not only to obtain a characterization of seeds from different genetic resources but also to detect human and/or technical errors that can eventually occur during the chain of the selfing cycles.
NOTE: Record the seed traits as detailed below at the beginning of each primary seed increase cycle. The phenotypic characterization must be performed before using the seeds in any (and all) selfing cycles.

Seeds Ruler
Analytical balance Spreadsheet software 1. Take at least five seeds from each accession (heterogeneous materials) or line (previously developed SSD lines), and through visual observation, evaluate and classify the Seed Coat Pattern according to the following categories (Fig. 4): • 8 = bipartite/tripartite • 9 = covered • 10 = coated • 11 = other (specify in the Notes).
2. Take at least five seeds from each accession (heterogeneous materials) or line (previously developed SSD lines), and through visual observation, evaluate and classify the Seed Coat Coloring according to the following three categories: • 1 = single colored • 2 = two colored • 3 = three (or more than three) colored.
3. Take at least five seeds from each accession (heterogeneous materials) or line (previously developed SSD lines), and through visual observation, evaluate and classify the Seed Coat Ground Color (i.e., primary color) according to the following 24 categories: • 5. Take at least five seeds from each accession (heterogeneous materials) or line (previously developed SSD lines), and through visual observation, evaluate the presence/absence of Seed Color Veining.
6. Take at least five seeds from each accession (heterogeneous materials) or line (previously developed SSD lines), and through visual observation, evaluate and classify the Seed Brilliance (i.e., seed shininess or opaqueness at harvest) according to one of the following three categories: 7. Take five seeds from each accession (heterogeneous materials) or line (previously developed SSD lines), and using a ruler, measure the Seed Coat Length (mm) by performing a lateral measurement, parallel to the hilum (Fig. 5). Use the mean of the values obtained from five seeds as the final measure.
8. Take five seeds from each accession (heterogeneous materials) or line (already developed SSD lines), and using a ruler, measure the Seed Coat Height (mm) by performing a lateral measurement, measured from the hilum to the opposite side (Fig. 5). Use the mean of the values obtained from five seeds as the final measure.
9. Take five seeds from each accession (heterogeneous materials) or line (already developed SSD lines), and using a ruler, measure the Seed Coat Width (mm) by Cortinovis et al.

of 28
Current Protocols Figure 5 Seed length, height, and width measurements. performing a ventral measurement (Fig. 5). Use the mean of the values obtained from five seeds as the final measure.
10. Take at least five seeds from each accession (heterogeneous materials) or line (already developed SSD lines), and through visual observation, evaluate and classify the Seed Shape according to one of the following five categories (Fig. 6): • 1 = round • 2 = oval • 3 = cuboid • 4 = kidney shaped • 5 = truncate fastigiate.
11. Measure the 100-Seed Weight, preferably by measuring the seed weight of at least two samples of 100 seeds (if not possible, then the seed weight should be measured on 10 seeds, as three different samples). Immature and/or infected seeds should be excluded.

BEAN SEED IMAGING
The aim of SEED bean imaging is to document the beans during their respective propagation steps in regard to the following traits: • Visual quality control of the propagation result; • Reference images for shape, color, and size; Cortinovis et al.

of 28
Current Protocols • Picture material for presenting the SSD line in various media; • Basis for automated image analysis.
A comprehensive discussion on the digitization of collections based on Nelson, Paul, Riccardi, and Mast (2012) is given in de la Hidalga, van Walsun, Rosin, Sun, and Wijers (2019). Accordingly, the following requirements should be fulfilled as the best-practice standard.
NOTE: Image capture should be done using a defined process under comparable conditions, and any institution and/or research group involved in seed increase of shared SSD lines should implement a standardized system to take images of seeds.
Important considerations for photography are listed below.
• To obtain good separation of the beans from the background, it is advisable to place them on a background of a color not present in skins of the beans. Light gray is ideal here for a neutral picture impression; if this is too close to the color of the beans, a black or white background can be used. The background should be flat and should not have any structural pattern. • Include a color chart in the photographs for quality control and postprocessing. This allows the lighting, white balance, and color accuracy of the image to be verified. • Include a scale bar to allow measurement of the dimensions of the objects.

Materials
Uniform background (see above) Color chart: e.g., ColorGauge Micro Target (Image Science Associates) or ColorChecker Classic Mini or Nano (X-Rite) Scale bar Photo lighting studio made of a lighting box (40 × 40 cm) with brightness-adjustable LED lamps (allowing gradual adjustment of light level) Microscope camera of at least 20 megapixels and an appropriate macro lens mounted on a stable tripod above the box Image-editing software for post-processing images that allows image normalization and resizing, as well as the subsequent addition of text or QR codes to the image Imaging process 1. Select at least five beans to be photographed. Selected bean seeds should be typical of the respective line in terms of color and size, although any variability should also be captured.
2. Place the bean seeds, color chart, and scale bar on the background. Arrange the beans so that they are separated from each other, and position one bean near the rule, at scale position zero of the scale. Be sure to leave enough space for the subsequent insertion of the label information and barcode.
3. Begin photography, processing only one bean line per pass to avoid mixing or confusion of the material. Before recording, verify the identity of the material.
4. Within the project space, the name of the image should be set to start with the unique IncreaseID followed by the sample ID or plant ID, and a unique image number separated by underscores, as: <IncreaseID>_<SampleID>_<ImageID> 5. Set the image resolution and size. The resolution and size of an image depends on the intended use. The higher the initial values, the more versatile the application is, as smaller versions can be easily created. The following values in Table 1 provide the guidelines for the resolution. The color bit depth must always be 24-bit color.

of 28
Current Protocols

Research and preservation 600
Printing 300 Web publishing 72 6. Process the images (e.g., image enhancement, normalization, addition of label information, barcode). For the barcode, different variants are possible.
a. Two-dimensional QR Code: This is preferred because further information can be stored in addition to the identifier. The data record for the examples given here is structured as follows: • Full Image-ID: <IncreaseID>_<SampleID>_<ImageID> • Image of: • DOI: <DOI of the SSD line that is shown> 8. Integrate the image identification and description information as a label applied subsequent to the image processing. The human-readable information is limited to the identifiers of the depicted object, and the reference to the image rights license (here, the Creative Commons license CC BY-SA, https:// creativecommons.org). Optionally, a logo can be included to boost brand recognizability.
An example of photo documentation of an SSD cycle 1 bean harvest is given in Figure  7. A comparison of SSD harvest with the appearance of bean seed characteristics of the corresponding gene bank accession is shown in Figure 8. The usefulness of the light gray background for bean seed color and pattern depiction is demonstrated in Figure 9.

CHARACTERIZING BEAN LINES FOR PLANT TRAIT DESCRIPTORS SPECIFIC FOR COMMON BEAN PRIMARY SEED INCREASE
The main aim of this protocol is the phenotypic evaluation of the different lines grown in controlled conditions during the cycles of seed multiplication. These phenotypic data, along with those obtained by applying Basic Protocols 1 and 2, will be uploaded into the project database and integrated with those obtained for the same lines in other experiments (i.e., field trials). All of these data will represent precious information for future users, such as researchers, breeders, and so on.  Users should record the following Mandatory Traits (Priority 1, steps 1-27), which are considered as essential within INCREASE and must be recorded for each selfing cycle, along with Non-Mandatory Traits (Priority 2, steps 28-33). Non-Mandatory Traits are considered within INCREASE as non-essential, but preferable. The collection of Non-Mandatory Traits overlaps with the collection of Mandatory Traits, and thus the step numbers are not fully sequential in time.
NOTE: This protocol assumes that plants and seeds are evaluated by visual inspection and manual measurement.
NOTE: During Primary Seed Increase, one plant is grown and characterized for each line, for at least the first two selfing cycles (development of SSD seeds); in subsequent cycles it may be possible to grow more than one plant per line, depending on the available space in controlled conditions (see Fig. 3). If more than three plants are grown per line, traits can be recorded on a subsample (at least three randomly chosen plants).
NOTE: Before starting the protocol, collect the following information on the experimental site: data collector name, location of experimental trial, latitude of Cortinovis et al.

of 28
Current Protocols Figure 9 Example of color compatibility with the chosen light gray background for beans. This allows color differences to be depicted, which can be compensated by normalization using the color chart.
experimental trial, longitude of experimental trial, altitude of experimental trial, and controlled-condition/insect-free measures applied (e.g., tunnel, greenhouse, grow chamber).

Seeds Ruler
Analytical balance Spreadsheet software

Recording Mandatory Traits (Priority 1)
1. Record the Sowing Date (i.e., date on which the seeds were sown).
2. Record the Days to Emergence (i.e., number of days after which the seedlings emerged, starting from the sowing day).
Emergence is defined here as the time at which seedling cotyledons/leaves become visible.
3. Through visual observation, evaluate and classify the Hypocotyl Pigmentation (i.e., color of the hypocotyl) according to one of the following three categories: • 1 = purple • 2 = green • 3 = other (specify in the Notes).

of 28
Current Protocols 4. Through visual observation, evaluate and classify the Leaf Color (chlorophyll) according to one of the following three categories of green: • 1 = pale green • 2 = medium green • 3 = dark green.  11. Through visual observation, evaluate and classify the Pod Curvature for fully expanded immature pods according to one of the following four categories (Fig. 11): • 1 = straight • 2 = slightly curved • 3 = curved • 4 = recurving.
12. Through visual observation, evaluate and classify the Pod Color at Physiological Maturity (PM) according to one of the following seven categories (see Fig. 12 for an example of bean pods at physiological maturity): •

of 28
Current Protocols Figure 11 Pod curvature patterns.

Figure 12
Bean plant bearing pods showing mature color (i.e., at physiological maturity).

of 28
Current Protocols Figure 13 Pod wall fiber patterns.
14. Through visual observation, evaluate and classify the Pod Wall Fiber of the mature pods according to one of the following three categories (Fig. 13): • 1 = strongly contracting (at dry maturity, adhering around seed) • 2 = leathery podded (dry pods will not spontaneously open) • 3 = excessive shattering (with strong twisting of dry pods). 25. Count and record the 100-Seed Mass (g).
Do not calculate this value from "Total seed mass" and "Total number of seeds." 26. Record Diseases, and if possible, describe or make notes regarding disease status.
There is a "Disease-specific comments" column for making any notes related to diseases, including but not limited to the observation that many or specific diseases are present.
27. Record the Stress Susceptibility, and if possible, describe or make notes about this.
There is a "Stress-specific comments" column for making any notes related to stress.
29. Record the Leaf Shape of the terminal leaflet of the third trifoliate leaf according to the following three categories (Fig. 14): • 1 = triangular • 2 = quadrangular • 3 = round.
30. Record the Pod Color for fully expanded immature pods according to the following seven categories: • 1 = light yellow • 2 = golden yellow/dark yellow • 3 = green/dark green • 4 = light green/gray-green • 5 = red • 6 = purple • 7 = other (specify in the Notes). 32. Record the Pod Suture String status on fully expanded immature pods from three randomly chosen pods, but only if the seed multiplication will be not compromised. Classify into one of the following four categories: • 0 = stringless • 1 = few strings • 2 = moderately stringy • 3 = very stringy.

Background Information
Here, we report an established system (i.e., phenotyping procedures and protocols) for the development of SSD lines starting from common bean genetic resources, for their maintenance and for the production of seeds for distribution. This system is based on a set of phenotyping protocols that can be applied by gene banks and research institutes that also want to use such resources. The strength of this system is that it is designed to produce both the genomic data associated with each pure line and an "open" platform for integration of data produced within INCREASE. This will also include data that are produced after the conclusion of INCREASE for the same materials. The idea is to set up a centralized system that can be used by gene banks and research institutes that can also integrate new data, and at the same time, be freely accessible by any user.