Into the wild: The soybean genome meets its undomesticated relative

Soybean (Glycine max) is one of the most widely grown crop species in the world. One of the major agricultural challenges of the 21st century will be to increase the yield of soybean and other major crop species to feed a growing population on a finite amount of farmland. Soybean breeding and improvement is hindered by a narrow domesticated germplasm relative to other crop species (1). Despite its importance, many outstanding questions remain regarding important aspects of soybean germplasm, including the extent of genomic variation within the domesticated germplasm and among domesticated and wild relatives. Glycine soja is the closest extant wild relative of soybean and is generally considered to be the undomesticated progenitor of the domesticated soybean. G. max and G. soja are phenotypically disparate in many ways, but they readily cross with one another and give rise to fertile hybrids, thus making G. soja a promising source of novel genes and alleles for soybean breeding and improvement.

The genome sequence of domesticated soybean was published earlier this year (2), bringing in a new era for soybean functional and comparative genomics. Comparative sequencing of soybean domesticates and wild relatives will substantially increase our understanding of the limitations of the domesticated germplasm and the potential to use wild relatives for crop breeding and improvement. The PNAS report by Kim et al. (3) focuses on the resequencing of wild soybean G. soja (accession no. IT182932) and the subsequent comparative genomic analysis with the reference G. max genome (2). Kim et al. (3) cataloged a wide range of nucleotide and structural variations between wild and domesticated soybean. A summary of the types and frequencies of the different gene variation classes identified in the Kim et al. analysis (3) are shown in Fig. 1. Nucleotide variants, such as base substitutions and small insertions and deletions, occurred at a frequency of 0.31% across the G. max and G. soja genomes. These types of alterations may affect the function and/or protein structure of more than 10,000 putative protein-encoding genes. Furthermore, many structural genomic differences were also apparent between G. max and G. soja, such as large insertions, deletions, and inversions. Deletions in the G. soja genome ranging from 100 bp to 100 kb may explain much of this structural variation. In total, approximately 1,000 genes were identified within regions of structural variation between G. max and G. soja. Although it is difficult to estimate what portion of these genes may in fact be located within respective sequence gaps, these results are corroborated by the recent resequencing of another G. soja accession (W05), which exhibited a similar number of gene content variants compared with G. max (4).

Fig. 1.

Nucleotide and structural variation identified between domesticated soybean (G. max) and wild soybean (G. soja) in gene coding regions. The arrows indicate gene positions along a hypothetical chromosomal region. The numbers of genes exhibiting variation between G. max and G. soja for each type of variation are shown in parentheses. Nucleotide variants that influence protein function or structure, such as base substitutions and small frame-shift mutations, are shown on the left (red triangles represent sites of nucleotide differences). Genomic structural variations, such as inversions, deletions, insertions, and translocations, are shown on the right. The number of genes that are found in G. soja and missing in G. max is ambiguous because of sequence gaps in the reference sequence; however, several examples of G. soja-specific genes were validated. The methodology used in this study (3) was unable to resolve chromosomal translocations, so the number of genes in this category remains unknown.

The exact timeline of soybean domestication remains a matter of dispute. Most estimates approximate that domestication occurred somewhere between 3,100 and 9,000 y ago (5, 6). Kim et al. (3) used their comparative sequence data to estimate the time of divergence between G. max and the G. soja accession they sequenced. Surprisingly, they estimated that the split occurred approximately 270,000 y ago, substantially predating soybean domestication. The authors concede that this may be an overestimate, as human selection may have increased the frequency of variation in seemingly neutral genes. However, the discrepancy between the timing of the G. max/G. soja split and the timing of domestication suggests that the domestication process may have been more complicated than has been thought, perhaps occurring in a lineage that split long ago from the G. soja accession sequenced by Kim et al. (3).

Furthermore, the comparative sequence data may be significant for understanding the genetic mechanisms of soybean domestication. The domestication syndrome that distinguishes G. max and G. soja is vast, including differences in plant architecture, flowering time, pod dehiscence, seed size, and other characteristics. Quantitative trait loci (QTLs) have been genetically mapped for several soybean domestication traits (7), but only one gene associated with domestication has been characterized to date (8). The G. soja sequence will be an important resource for identifying candidate domestication genes. This type of analysis will be particularly powerful when combined with population-level comparative sequencing, allowing for the identification of regions of conserved divergence between the domesticated and wild accessions (4).

Perhaps the most important use of the G. soja sequence will be to identify genes and alleles from the wild germplasm that may have potential applications for use in soybean cultivar improvement. Several major crop species, including tomato, barley, and wheat, have made substantial use of their wild relatives to expand their gene pools and incorporate novel traits, particularly pest and disease resistance (9–11). Soybean breeding has had less success at incorporating wild introgressions into elite cultivars for a variety of reasons (5). However, the influence of wild introgressions on the soybean germplasm may be underestimated, as recent studies indicated that G. soja introgressions are found in some soybean accessions and breeding lines (4, 12).

The allelic diversity in G. soja is greater than that of soybean (4). The sequence comparisons by Kim et al. (3) and Lam et al. (4) have identified a subset of genes and alleles in G. soja that are not found in the soybean reference sequence. Efforts have been and are being made to assess the impact of wild G. soja genetic introgressions on soybean phenotypes; in fact, the soybean breeding community has identified several QTLs for which the G. soja locus is more favorable than the G. max locus for specific qualitative and quantitative traits of interest (13–18). Linking the molecular sequence variation to the phenotypic variation between G. max and G. soja is clearly the next challenge. From a practical standpoint, the G. max/G. soja nucleotide and structural diversity revealed by Kim et al. (3) will be useful as a reference for the genetic mapping of G. soja introgressions in soybean populations and the identification of novel candidate genes and alleles in the G. soja sequence that may underlie QTLs conferring superior phenotypes. The identification of structural variants may also be useful for understanding why certain regions of the G. soja genome may be recalcitrant to stable introgression.

At present, it is clear that a revolution in the comparative sequencing of major crop species, like maize (19), rice (20), and soybean (3, 4), is well under way. The number of sequenced accessions available to the public will continue to grow rapidly, offering new opportunities and challenges for breeders, population geneticists, and molecular biologists. It is easy to recognize the potential of these resources from a research standpoint, but the real challenge may be in translating this new knowledge into advances in crop productivity and stability.