BackgroundSequencing of cDNA libraries for the development of expressed sequence tags (ESTs) as well as for the discovery of simple sequence repeats (SSRs) has been a common method of developing microsatellites or SSR-based markers. In this research, our objective was to further sequence and develop common bean microsatellites from leaf and root cDNA libraries derived from the Andean gene pool accession G19833 and the Mesoamerican gene pool accession DOR364, mapping parents of a commonly used reference map. The root libraries were made from high and low phosphorus treated plants.ResultsA total of 3,123 EST sequences from leaf and root cDNA libraries were screened and used for direct simple sequence repeat discovery. From these EST sequences we found 184 microsatellites; the majority containing tri-nucleotide motifs, many of which were GC rich (ACC, AGC and AGG in particular). Di-nucleotide motif microsatellites were about half as common as the tri-nucleotide motif microsatellites but most of these were AGn microsatellites with a moderate number of ATn microsatellites in root ESTs followed by few ACn and no GCn microsatellites. Out of the 184 new SSR loci, 120 new microsatellite markers were developed in the BMc (Bean Microsatellites from cDNAs) series and these were evaluated for their capacity to distinguish bean diversity in a germplasm panel of 18 genotypes. We developed a database with images of the microsatellites and their polymorphism information content (PIC), which averaged 0.310 for polymorphic markers.ConclusionsThe present study produced information about microsatellite frequency in root and leaf tissues of two important genotypes for common bean genomics: namely G19833, the Andean genotype selected for whole genome shotgun sequencing from race Peru, and DOR364 a race Mesoamerica subgroup 2 genotype that is a small-red seeded, released variety in Central America. Both race Peru and Mesoamerica subgroup 2 (small red beans) have been understudied in comparison to race Nueva Granada and Mesoamerica subgroup 1 (black beans) both with regards to gene expression and as sources of markers. However, we found few differences between SSR type and frequency between the G19833 leaf and DOR364 root tissue-derived ESTs. Overall, our work adds to the analysis of microsatellite frequency evaluation for common bean and provides a new set of 120 BMc markers which combined with the 248 previously developed BMc markers brings the total in this series to 368 markers. Once we include BMd markers, which are derived from GenBank sequences, the current total of gene-based markers from our laboratory surpasses 500 markers. These markers are basic for studies of the transcriptome of common bean and can form anchor points for genetic mapping studies in the future.

Microsatellite markers are useful genetic tools for a wide array of genomic analyses although their development is time-consuming and requires the identification of simple sequence repeats (SSRs) from genomic sequences. Screening of non-enriched, small-insert libraries is an effective method of SSR isolation that can give an unbiased picture of motif frequency. Here we adapt high-throughput protocols for the screening of plasmid-based libraries using robotic colony picking and filter preparation. Seven non-enriched genomic libraries from common bean genomic DNA were made by digestion with four frequently cutting restriction enzymes, double digestion with a frequently cutting restriction enzyme and a less frequently cutting restriction enzyme, or sonication. Library quality was compared and three of the small-insert libraries were selected for further analysis. Each library was plated and picked into 384-well plates that were used to create high-density filter arrays of over 18 000 clones each, which were screened with oligonucleotide probes for various SSR motifs. Positive clones were found to have low redundancy. One hundred SSR markers were developed and 80 were tested for polymorphism in a standard parental survey. These microsatellite markers derived from non-SSR-enriched libraries should be useful additions to previous markers developed from enriched libraries.

Common bean is an important and diverse crop legume with several wild relatives that are all part of the Phaseoleae tribe of tropical crop legumes. Sequence databases have been a good source of sequences to mine for simple sequence repeats (SSRs). The objective of this research was to evaluate 14 sequence collections from common bean for SSRs and to evaluate the diversity of the polymorphic microsatellites derived from these collections. SSRs were found in 10 of the GenBank sequence collections with an average of 11.3% of sequences containing microsatellite motifs. The most common motifs were based on tri- and dinucleotides. In a marker development programme, primers were designed for 125 microsatellites which were tested on a panel of 18 common bean genotypes. The markers were named as part of the bean microsatellite-database (BMd) series, and the average polymorphism information content was 0.404 for polymorphic markers and predicted well the genepool structure of common beans and the status of the wild and cultivated accessions that were included in the study. Therefore, the BMd series of microsatellites is useful for multiple studies of genetic relatedness and as anchor markers in future mapping of wide crosses in the species.

Expressed sequence tags (ESTs) are a rich source of SSR sequences, but the proportion of long Class I microsatellites with many repeats vs. short Class II microsatellites with few repeats is an important factor to consider. Class I microsatellites, with more than 20 bp of repeats, tend to make better markers with higher polymorphism. The goal of this study was to determine the frequency of Class I and Class II microsatellites in a collection of over 21,000 ESTs from a single study of five different tissues of common bean: two types of leaves, nodules, pods and roots. For this objective, we used three different bioinformatics pipelines: Automated Microsatellite Marker Development (AMMD), Batchprimer3 and SSRLocator. In addition, we determined the frequency of single or multiple SSRs in the assembled ESTs, the frequency of perfect and compound repeats and whether Class I microsatellites were mainly di-nucleotide or tri-nucleotide motifs with each of the search engines. Primers were designed for a total of 175 microsatellites concentrating on class I microsatellites identified with SSR locator. A few other microsatellites were included from the other search engines, AMMD and Batchprimer3 programs so as to have a representative set of class II markers for comparison sake. The comparison of 95 class I vs. 80 class II markers confirmed that the Class I were more polymorphic and therefore more useful.