QTL for seed protein and amino acids in the Benning × Danbaekkong soybean population.

KEY MESSAGE: We identified QTL associated with protein and amino acids in a soybean mapping population that was grown in five environments. These QTL could be used in MAS to improve these traits. Soybean, rather than nitrogen-containing forages, is the primary source of quality protein in feed formulations for domestic swine, poultry, and dairy industries. As a sole dietary source of protein, soybean is deficient in the amino acids lysine (Lys), threonine (Thr), methionine (Met), and cysteine (Cys). Increasing these amino acids would benefit the feed industry. The objective of the present study was to identify quantitative trait loci (QTL) associated with crude protein (cp) and amino acids in the 'Benning' × 'Danbaekkong' population. The population was grown in five southern USA environments. Amino acid concentrations as a fraction of cp (Lys/cp, Thr/cp, Met/cp, Cys/cp, and Met + Cys/cp) were determined by near-infrared reflectance spectroscopy. Four QTL associated with the variation in crude protein were detected on chromosomes (Chr) 14, 15, 17, and 20, of which, a QTL on Chr 20 explained 55 % of the phenotypic variation. In the same chromosomal region, QTL for Lys/cp, Thr/cp, Met/cp, Cys/cp and Met + Cys/cp were detected. At these QTL, the Danbaekkong allele resulted in reduced levels of these amino acids and increased protein concentration. Two additional QTL for Lys/cp were detected on Chr 08 and 20, and three QTL for Thr/cp on Chr 01, 09, and 17. Three QTL were identified on Chr 06, 09 and 10 for Met/cp, and one QTL was found for Cys/cp on Chr 10. The study provides information concerning the relationship between crude protein and levels of essential amino acids and may allow for the improvement of these traits in soybean using marker-assisted selection.

BARCSoySNP23: a panel of 23 selected SNPs for soybean cultivar identification.

This report describes a set of 23 informative SNPs (BARCSoySNP23) distributed on 19 of the 20 soybean linkage groups that can be used for soybean cultivar identification. Selection of the SNPs to include in this set was made based upon the information provided by each SNP for distinguishing a diverse set of soybean genotypes as well as the linkage map position of each SNP. The genotypes included the ancestors of North American cultivars, modern North American cultivars and a group of Korean cultivars. The procedure used to identify this subset of highly informative SNP markers resulted in a significant increase in the power of identification versus any other randomly selected set of equal number. This conclusion was supported by a simulation which indicated that the 23-SNP panel can uniquely distinguish 2,200 soybean cultivars, whereas sets of randomly selected 23-SNP panels allowed the unique identification of only about 50 cultivars. The 23-SNP panel can efficiently distinguish each of the genotypes within four maturity group sets of additional cultivars/lines that have identical classical pigmentation and morphological traits. Comparatively, the 13 trinucleotide SSR set published earlier (BARCSoySSR13) has more power on a per locus basis because of the multi-allelic nature of SSRs. However, the assay of bi-allelic SNP loci can be multi-plexed using non-gel based techniques allowing for rapid determination of the SNP alleles present in soybean genotypes, thereby compensating for their relatively low information content. Both BARCSoySNP23 and BARCSoySSR13 were highly congruent relative to identifying genotypes and for estimating population genetic differences.

Discovery of SNPs in soybean genotypes frequently used as the parents of mapping populations in the United States and Korea.

Single nucleotide polymorphisms (SNPs) including insertion/deletions (indels) serve as useful and informative genetic markers. The availability of high-throughput and inexpensive SNP typing systems has increased interest in the development of SNP markers. After fragments of genes were amplified with primers derived from 110 soybean GenBank ESTs, sequencing data of PCR products from 15 soybean genotypes from Korea and the United States were analyzed by SeqScape software to find SNPs. Among 35 gene fragments with at least one SNP among the 15 genotypes, SNPs occurred at a frequency of 1 per 2,038 bp in 16,302 bp of coding sequence and 1 per 191 bp in 16,960 bp of noncoding regions. This corresponds to a nucleotide diversity (theta) of 0.00017 and 0.00186, respectively. Of the 97 SNPs discovered, 78 or 80.4% were present in the six North American soybean mapping parents. The addition of "Hwaeomputkong," which originated from Japan, increased the number to 92, or 94.8% of the total number of SNPs present among the 15 genotypes. Thus, Hwaeomputkong and the six North American mapping parents provide a diverse set of soybean genotypes that can be successfully used for SNP discovery in coding DNA and closely associated introns and untranslated regions.

Development and mapping of microsatellite (SSR) markers in wheat.

Microsatellite DNA markers are consistently found to be more informative than other classes of markers in hexaploid wheat. The objectives of this research were to develop new primers flanking wheat microsatellites and to position the associated loci on the wheat genome map by genetic linkage mapping in the ITMI W7984 x Opata85 recombinant inbred line (RIL) population and/or by physical mapping with cytogenetic stocks. We observed that the efficiency of marker development could be increased in wheat by creating libraries from sheared rather than enzyme-digested DNA fragments for microsatellite screening, by focusing on microsatellites with the [ATT/TAA]n motif, and by adding an untemplated G-C clamp to the 5'-end of primers. A total of 540 microsatellite-flanking primer pairs were developed, tested, and annotated from random genomic libraries. Primer pairs and associated loci were assigned identifiers prefixed with BARC (the acronym for the USDA-ARS Beltsville Agricultural Research Center) or Xbarc, respectively. A subset of 315 primer sets was used to map 347 loci. One hundred and twenty-five loci were localized by physical mapping alone. Of the 222 loci mapped with the ITMI population, 126 were also physically mapped. Considering all mapped loci, 126, 125, and 96 mapped to the A, B, and D genomes, respectively. Twenty-three of the new loci were positioned in gaps larger than 10 cM in the map based on pre-existing markers, and 14 mapped to the ends of chromosomes. The length of the linkage map was extended by 80.7 cM. Map positions were consistent for 111 of the 126 loci positioned by both genetic and physical mapping. The majority of the 15 discrepancies between genetic and physical mapping involved chromosome group 5.

Comparison of four flow cytometric SNP detection assays and their use in plant improvement.

Single nucleotide polymorphisms (SNPs) are attractive DNA markers due to their abundance and potential for use in automated high-throughput genotyping. Numerous SNP genotyping assays have been developed, but it is unclear which assays are best suited and most efficient for various types of plant improvement research. The objective of this study was to compare the accuracy, efficiency, and cost of four SNP genotyping assays: single-base extension (SBE), allele-specific primer extension (ASPE), oligonucleotide ligation (OL), and direct hybridization (DH). All four assay methods used the same Luminex 100 flow cytometer platform. Fifty-eight F(2)-derived soybean [Glycine max (L.) Merr.] lines from a cross between inbred lines G99-G725 and N00-3350 were genotyped at four SNPs. SBE and ASPE clearly differentiated between the two homozygotes and the heterozygote at each SNP. Results were in agreement with those identified using the SNaPshot minisequencing assay as a control. In contrast, the OL and DH assays were unable to differentiate between genotypes at some of the SNPs. However, when the cost per data point for the four different assays was compared, the cost of OL and DH was only about 70% of that for SBE, with DH requiring the least time of the four assays. On the basis of cost and labor, ASPE is more cost-effective and simpler than SBE, and would therefore be a good method for genetic mapping and diversity studies which require a large number of markers and a high level of multiplexing. DH appears to be the most economical assay for marker-assisted selection, though optimization for DH would be required for some SNP markers.

A new integrated genetic linkage map of the soybean.

A total of 391 simple sequence repeat (SSR) markers designed from genomic DNA libraries, 24 derived from existing GenBank genes or ESTs, and five derived from bacterial artificial chromosome (BAC) end sequences were developed. In contrast to SSRs derived from EST sequences, those derived from genomic libraries were a superior source of polymorphic markers, given that the mean number of tandem repeats in the former was significantly less than that of the latter ( P<0.01). The 420 newly developed SSRs were mapped in one or more of five soybean mapping populations: "Minsoy" x "Noir 1", "Minsoy" x "Archer", "Archer" x "Noir 1", "Clark" x "Harosoy", and A81-356022 x PI468916. The JoinMap software package was used to combine the five maps into an integrated genetic map spanning 2,523.6 cM of Kosambi map distance across 20 linkage groups that contained 1,849 markers, including 1,015 SSRs, 709 RFLPs, 73 RAPDs, 24 classical traits, six AFLPs, ten isozymes, and 12 others. The number of new SSR markers added to each linkage group ranged from 12 to 29. In the integrated map, the ratio of SSR marker number to linkage group map distance did not differ among 18 of the 20 linkage groups; however, the SSRs were not uniformly spaced over a linkage group, clusters of SSRs with very limited recombination were frequently present. These clusters of SSRs may be indicative of gene-rich regions of soybean, as has been suggested by a number of recent studies, indicating the significant association of genes and SSRs. Development of SSR markers from map-referenced BAC clones was a very effective means of targeting markers to marker-scarce positions in the genome.

Assignment of molecular linkage groups to soybean chromosomes by primary trisomics.

Gene-linkage groups (classical linkage groups, CLGs; molecular linkage groups, MLGs) and chromosome relationship in soybean [ Glycine max (L.) Merr., 2n = 40] is not yet established. However, primary trisomics provide an invaluable cytogenetic tool to associate genes and linkage groups to specific chromosomes. We have assigned 11 MLGs to soybean chromosomes by using primary trisomics (2 x + 1 = 41) and SSR markers. Primary trisomics were hybridized with Glycine soja Sieb. and Zucc. (2n = 40) in the greenhouse, F(1) plants with 2n = 40 and 41 were identified cytologically and 41 chromosome plants were selfed. A deviation from the 1:2:1 ratio in the F(2) population suggests a marker is associated with a chromosome. Of the possible 220 combinations involving 20 MLGs and 11 primary trisomics, 151 combinations were examined. The relationships between soybean chromosomes and MLGs are: 1 = D1a+q, 3 = N, 5 = A1, 8 = A2, 9 = K, 13 = F, 14 = C1, 17 = D2, 18 = G, 19 = L and 20 = I. This study sets the stage to establish relationship between nine remaining MLGs with the other genetically unidentified nine primary trisomics. The association of CLGs with the soybean chromosomes will be discussed.

Single-nucleotide polymorphisms in soybean.

Single-nucleotide polymorphisms (SNPs) provide an abundant source of DNA polymorphisms in a number of eukaryotic species. Information on the frequency, nature, and distribution of SNPs in plant genomes is limited. Thus, our objectives were (1) to determine SNP frequency in coding and noncoding soybean (Glycine max L. Merr.) DNA sequence amplified from genomic DNA using PCR primers designed to complete genes, cDNAs, and random genomic sequence; (2) to characterize haplotype variation in these sequences; and (3) to provide initial estimates of linkage disequilibrium (LD) in soybean. Approximately 28.7 kbp of coding sequence, 37.9 kbp of noncoding perigenic DNA, and 9.7 kbp of random noncoding genomic DNA were sequenced in each of 25 diverse soybean genotypes. Over the >76 kbp, mean nucleotide diversity expressed as Watterson's theta was 0.00097. Nucleotide diversity was 0.00053 and 0.00111 in coding and in noncoding perigenic DNA, respectively, lower than estimates in the autogamous model species Arabidopsis thaliana. Haplotype analysis of SNP-containing fragments revealed a deficiency of haplotypes vs. the number that would be anticipated at linkage equilibrium. In 49 fragments with three or more SNPs, five haplotypes were present in one fragment while four or less were present in the remaining 48, thereby supporting the suggestion of relatively limited genetic variation in cultivated soybean. Squared allele-frequency correlations (r(2)) among haplotypes at 54 loci with two or more SNPs indicated low genome-wide LD. The low level of LD and the limited haplotype diversity suggested that the genome of any given soybean accession is a mosaic of three or four haplotypes. To facilitate SNP discovery and the development of a transcript map, subsets of four to six diverse genotypes, whose sequence analysis would permit the discovery of at least 75% of all SNPs present in the 25 genotypes as well as 90% of the common (frequency >0.10) SNPs, were identified.

Characterization of trinucleotide SSR motifs in wheat.

Length differences among trinucleotide-based microsatellite alleles can be more easily detected and frequently produce fewer "stutter bands" as compared to dinucleotide-based microsatellite markers. Our objective was to determine which trinucleotide motif(s) would be the most-polymorphic and abundant source of trinucleotide microsatellite markers in wheat ( Triticum aestivumL.). Four genomic libraries of cultivar 'Chinese Spring' were screened with nine trinucleotide probes. Based on the screening of 28550 clones, the occurrences of (CTT/GAA) (n), (GGA/CCT) (n), (TAA/ATT) (n), (CAA/GTT) (n), (GGT/CCA) (n), (CAT/GTA) (n), (CGA/GCT) (n), (CTA/GAT) (n), and (CGT/GCA) (n) repeats were estimated to be 5.4x10(4), 3.5x10(4), 3.2x10(4), 1.2x10(4), 6.3x10(3), 4.9x10(3), 4.5x10(3), 4.5x10(3) and 3.6x10(3), i.e., once every 293 kbp, 456 kbp, 500 kbp, 1.3 Mbp, 2.6 Mbp, 3.2 Mbp, 3.6 Mbp, 3.6 Mbp and 4.5 Mbp in the wheat genome, respectively. Of 236 clones selected for sequencing, 38 (93%) (TAA/ATT) (n), 30 (43%) (CTT/GAA) (n), 16 (59%) (CAA/GTT) (n), 3 (27%) (CAT/GTA) (n) and 2 (4%) (GGA/CCT) (n) clones contained microsatellites with eight or more perfect repeats. From these data, 29, 27 and 16 PCR primer sets were designed and tested to the (TAA/ATT) (n), (CTT/GAA) (n) and (CAA/GTT) (n) microsatellites, respectively. A total of 12 (41.4%) primers designed to (TAA/ATT) (n), four (14.8%) to (CTT/GAA) (n), and two (12.5%) to (CAA/GTT) (n) resulted in polymorphic markers. The results indicated that (TAA/ATT) (n) microsatellites would provide the most-abundant and the most-polymorphic source of trinucleotide microsatellite markers in wheat.

Simple sequence repeat DNA markers in alfalfa and perennial and annual Medicago species.

Simple sequence repeat (SSR) or microsatellite DNA markers have been shown to function well in plant and mammalian species for genetic map construction and genotype identification. The objectives of the work reported here were to search GenBank for the presence of SSR-containing sequences from the genus Medicago, to assess the presence and frequency of SSR DNA in the alfalfa (Medicago sativa (L.) L. &L.) genome, and to examine the function of selected markers in a spectrum of perennial and annual Medicago species. The screening of an alfalfa genomic DNA library and sequencing of clones putatively containing SSRs indicated approximately 19 000 (AT)n + (CT)n + (CA)n + (ATT)n SSRs in the tetraploid genome. Inheritance was consistent with Mendelian expectations at four selected SSR loci with different core motifs. Additionally, genotypes of a range of Medicago species, including 10 perennial subspecies of the M. sativa complex and other perennial and annual Medicago species, were analyzed at each of the loci to ascertain the presence, number, and size of SSR alleles at each locus in each genotype. These studies indicate that SSR markers can function in alfalfa for the construction of genetic maps and will also be useful in a range of Medicago species for purposes of assessing genetic relatedness and taxonomic relationships, and for genotype identification.

Length polymorphism and homologies of microsatellites in several Cucurbitaceae species.

The objectives of this research were to assess (1) the degree of Simple Sequence Repeats (SSR) DNA length polymorphism in melon (Cucumis melo L.) and other species within the Cucurbitaceae family and (2) the possibility of utilizing SSRs flanking primers from single species to other genera or species of Cucurbitaceae. Five melon (CT/GA) n SSRs were isolated from a genomic library. Two cucumber (Cucumis sativus L.) SSRs were detected through a search of DNA sequence databases, one contained a (CT)8 repeat, the other a (AT)13 repeat. The seven SSRs were used to test a diverse sample of Cucurbitaceae, including 8 melon, 11 cucumber, 5 squash, 1 pumpkin, and 3 watermelon genotypes. Five of the seven SSRs detected length polymorphism among the 8 melon genotypes. PCR amplification revealed between three and five length variants (alleles) for each SSR locus, with gene diversity values ranging from 0.53 to 0.75. Codominant segregation of the alleles among F2 progeny was demonstrated for each of the five SSR loci. Four of the seven SSRs detected polymorphism among the 11 cucumber genotypes, with gene diversity values ranging between 0.18 and 0.64. Primers specific to SSRs of C. melo and C. sativus also amplified DNA extracted from genotypes belonging to other genera of the Cucurbitaceae family.

Restriction of Nodulation by Bradyrhizobium japonicum Is Mediated by Factors Present in the Roots of Glycine max.

Reciprocal grafting experiments done using soybean plant introduction genotypes indicated that restriction of nodulation by Bradyrhizobium japonicum is determined by the genotype of the root and is dependent on plant growth temperature. Microscopic analyses indicated that the soybean plant introduction genotypes restrict nodulation of B. japonicum at symbiotic stages which occur both before and after the formation of nodule primordia.

The analysis of simple sequence repeat DNA in soybean by Capillary Gel Electrophoresis.

The objective of this work is to examine the presence of simple sequence repeat (SSR) DNA in soybean plant genotypes by Capillary Gel Electrophoresis (CGE). The SSR DNA length polymorphism in soybean determines the variation in polymerase chain reaction (PCR) product lengths. Loci were chosen where amplification produced one PCR product per genotype (M.S. Akkaya et al., 1992). The F1 hybrids of parents carrying different alleles produced two PCR products identical to the two parents. The CGE system used a 3%T,3%C polyacrylamide gel capillary with an effective length of 40 cm. The PCR products with lengths of 150 to 200 base pairs were monitored at 260 nm. The analysis time was under 50 minutes. CGE is capable of separating these PCR products by base pair number the same as conventional sequencing gel techniques. CGE offers an automated, high speed, high resolution analytical method for determining soybean SSR allele sizes as compared with the traditional methodologies.

The use of microsatellite DNA markers for soybean genotype identification.

Conventional morphological and pigementation traits, as well as disease resistance, have been used to distinguish the uniqueness of new soybean cultivars for purposes of plant variety protection. With increasing numbers of cultivars and a finite number of conventional characters, it has become apparent that such traits will not suffice to establish uniqueness. The objective of this work was to provide an initial evaluation of microsatellite or simple-sequence-repeat (SSR) DNA markers to develop unique DNA profiles of soybean genotypes. Microsatellites are DNA sequences such as (AT) n /(TA) n and (ATT) n /(TAA) n that are composed of tandemly repeated 2-5-basepair DNA core sequences. The DNA sequences flanking microsatellites are generally conserved allowing the selection of polymerase chain reaction (PCR) primers that will amplify the intervening SSR. Variation in the number of tandem repeats, "n", results in PCR product length differences. The SSR alleles present at three (AT) n /(TA) n and four (ATT) n /(TAA) n loci were determined in each of 96 diverse soybean genotypes. Between 11 and 26 alleles were found at each of the seven loci. Only two genotypes had identical SSR allelic profiles and these had very similar pedigrees. The gene diversity for the seven markers averaged 0.87 for all 96 genotypes and 0.74 for a subset of 26 North American cultivars. These are much higher than soybean gene diversity values obtained using RFLP markers, and are similar to the average values obtained for human microsatellite markers. SSR markers provide an excellent complement to the conventional markers that are currently used to characterize soybean genotypes.

Length polymorphisms of simple sequence repeat DNA in soybean.

The objective of this work was to ascertain the presence and degree of simple sequence repeat (SSR) DNA length polymorphism in the soybean [Glycine max (L.) Merr.]. A search of GenBank revealed no (CA)n or (GT)n SSRs with n greater than 8 in soybean. In contrast, 5 (AT)n and 1 (ATT)n SSRs with n ranging from 14 to 27 were detected. Polymerase chain reaction (PCR) primers to regions flanking the six SSR loci were used in PCR amplification of DNA from 43 homozygous soybean genotypes. At three loci, amplification produced one PCR product per genotype and revealed 6, 7 and 8 product length variants (alleles) at the three loci, respectively. F1 hybrids between parents carrying different alleles produced two PCR products identical to the two parents. Codominant segregation of alleles among F2 progeny was demonstrated at each locus. A soybean DNA library was screened for the presence of (CA/GT)n SSRs. Sequencing of positive clones revealed that the longest such SSR was (CA)9. Thus, (CA)n SSRs with n of 15 or more are apparently much less common in soybean than in the human genome. In contrast to humans, (CA)n SSRs will probably not provide an abundant source of genetic markers in soybean. However, the apparent abundance of long (AT)n sequences should allow this SSR to serve as a source of highly polymorphic genetic markers in soybean.

Hyperreiterated DNA regions are conserved among Bradyrhizobium japonicum serocluster 123 strains.

We have identified and cloned two DNA regions which are highly reiterated in Bradyrhizobium japonicum serocluster 123 strains. While one of the reiterated DNA regions, pFR2503, is closely linked to the B. japonicum common and genotype-specific nodulation genes in strain USDA 424, the other, pMAP9, is located next to a Tn5 insertion site in a host-range extension mutant of B. japonicum USDA 438. The DNA cloned in pFR2503 and pMAP9 are reiterated 18 to 21 times, respectively, in the genomes of B. japonicum serocluster 123 strains. Gene probes from the reiterated regions share sequence homology, failed to hybridize (or hybridized poorly) to genomic DNA from other B. japonicum and Bradyrhizobium spp. strains, and did not hybridize to DNA from Rhizobium meliloti, Rhizobium fredii, Rhizobium leguminosarum biovars trifolii, phaseoli, and viceae, or Agrobacterium tumefacians. The restriction fragment length polymorphism hybridization profiles obtained by using these gene probes are useful for discriminating among serologically related B. japonicum serocluster 123 strains.

The Soybean Rj4 Allele Restricts Nodulation by Bradyrhizobium japonicum Serogroup 123 Strains.

Of nine Bradyrhizobium japonicum serogroup 123 strains examined, 44% were found to be restricted for nodulation by cultivar Hill. Nodulation studies with soybean isoline BARC-2 confirmed that the soybean Rj4 allele restricts nodulation by the same serogroup 123 isolates. Immunological analyses indicated that B. japonicum strains in serogroups 123 and 31 share at least one surface somatic antigen.

The Bradyrhizobium japonicum nolA gene and its involvement in the genotype-specific nodulation of soybeans.

Several soybean genotypes have been identified which specifically exclude nodulation by members of Bradyrhizobium japonicum serocluster 123. We have identified and sequenced a DNA region from B. japonicum strain USDA 110 which is involved in genotype-specific nodulation of soybeans. This 2.3-kilobase region, cloned in pMJS12, allows B. japonicum serocluster 123 isolates to form nodules on plants of serogroup 123-restricting genotypes. The nodules, however, were ineffective for symbiotic nitrogen fixation. The nodulation-complementing region is located approximately 590 base pairs transcriptionally downstream from nodD2. The 5' end of pMJS12 contains a putative open reading frame (ORF) of 710 base pairs, termed nolA. Transposon Tn3-HoHo mutations only within the ORF abolished nodulation complementation. The N terminus of the predicted nolA gene product has strong similarity with the N terminus of MerR, the regulator of mercury resistance genes. Translational lacZ fusion experiments indicated that nolA was moderately induced by soybean seed extract and the isoflavone genistein. Restriction fragments that hybridize to pMJS12 were detected in genomic DNAs from both nodulation-restricted and -unrestricted strains.

DNA Hybridization Probe for Use in Determining Restricted Nodulation among Bradyrhizobium japonicum Serocluster 123 Field Isolates.

Several soybean plant introduction (PI) genotypes have recently been described which restrict nodulation of Bradyrhizobium japonicum serocluster 123 in an apparently serogroup-specific manner. While PI 371607 restricts nodulation of strains in serogroup 123 and some in serogroup 127, those in serogroup 129 are not restricted. When DNA regions within and around the B. japonicum I-110 common nodulation genes were used as probes to genomic DNA from the serogroup strains USDA 123, USDA 127, and USDA 129, several of the probes differentially hybridized to the nodulation-restricted and -unrestricted strains. One of the gene regions, cloned in plasmid pMJS12, was subsequently shown to hybridize to 4.6-kilobase EcoRI fragments from DNAs from nodulation-restricted strains and to larger fragments in nodulation-unrestricted strains. To determine if the different hybridization patterns could be used to predict nodulation restriction, we hybridized pMJS12 to EcoRI-digested genomic DNAs from uncharacterized serocluster 123 field isolates. Of the 36 strains examined, 15 were found to have single, major, 4.6-kilobase hybridizing EcoRI fragments. When tested for nodulation, 80% (12 of 15) of the strains were correctly predicted to be restricted for nodulation of the PI genotypes. In addition, hybridization patterns obtained with pMJS12 and nodulation phenotypes on PI 371607 indicated that there are at least three types of serogroup 127 strains. Our results suggest that the pMJS12 gene probe may be useful in selecting compatible host-strain combinations and in determining the suitability of field sites for the placement of soybean genotypes containing restrictive nodulation alleles.