Chapter 1
Genome Mapping and Agriculture
Randy C. Shoemaker1,2,3, Lisa L. Lorenzen3, Brian W. Diers4, and Terry C. Olson2
1USDA-ARS-FCR, 2Department of Agronomy, 3Department of Zoology/Genetics, Agronomy Hall, Iowa State University, Ames, IA 50011, USA, 4 Department of Crop and Soil Science, Michigan State University, East Lansing, MI 48824-1325, USA.
The advent of restriction endonuclease technologies and our ability to manipulate DNA have rapidly developed into a methodology that permeates every discipline involving the life sciences. Genotypic analysis, including genome mapping are emerging as important contributors to crop improvement. The purpose of this paper is to review progress made in these areas and to discuss the various potential applications of these methodologies for soybean improvement.
The soybean genome is approximately 40 - 60% repetitive sequences (Gurley et al, 1979; Goldberg, 1978). The majority (65 - 70%) of single-copy sequences have a short period interspersion with single-copy sequences of 1.1 - 1.4 kb alternating with repetitive sequence elements of 0.3 - 0.4 kb (Gurley et al, 1979). Analysis of pachytene chromosomes has shown that over 35% of the soybean genome is made up of heterochromatin--the short arms of six of the 20 bivalents seem to be completely heterochromatic (Singh and Hymowitz, 1988).
In order to adequately and accurately analyze band segregation patterns, low-or single-copy clones are required for use as probes. A high percentage of plant DNA is modified by methylation of the cytosine base. Many restriction endonucleases will not cleave DNA containing methylated cytosines within certain sequences. Keim and Shoemaker (1988) used the restriction endonuclease Pstl, a methylation-sensitive enzyme to construct a genomic library. Sequences in and adjacent to transcribed regions generally are unmethylated and therefore will be cleaved by this enzyme (Burr et al, 1988). Using this technique, a library of size-fractionated recombinant DNA was developed that was 80-85% single-copy DNA sequence (Keim and Shoemaker, 1988; Keim et al, 1990a).
When tested against DNA from the G. max breeding line A81-356022 and G. soja accession PI 468.916, approximately 40% of the random genomic probes detect polymorphisms with two or more enzymes, suggesting that DNA rearrangements are the cause of the polymorphisms. This hypothesis was substantiated through RFLP mapping of polymorphic regions (Apuya et al, 1988). Approximately 10% of the probes from our genome library detect “dominant” markers, i.e., the heterozygote classes are indistinguishable from dominant classes. These “dominant” markers could be the result of additions or deletions within one of the genotypes.
The molecular map was generated by analysis of marker segregation among 58 F2 individuals. The computer program “MapMaker” (Lander et al, 1987) was used to create the “best” loci order. A minimum LOD score of 3.0 was used for the pairwise linkage analysis in all instances. Nearly 450 qualitative markers have been placed on this map. These markers include random genomic clones, clones of known genes (cDNAs), isoenzyme loci, and a variety of morphological and developmental classical markers. This map includes 20 linkage groups of three or more markers and four linkage groups containing only two-point linkages. The linkage map encompasses approximately 3000 cM.
Integrating the molecular and classical map
Currently, the classical soybean genetic linkage map contains only 49 linked markers which covers approximately 530 map units (Palmer and Kiang, 1990). Hundreds of other known qualitative traits remain unmapped.
To exploit fully the potential of a molecular genetic map, it is necessary to integrate molecular and conventional markers into a unified linkage map. This can be accomplished through painstaking segregation and linkage analysis in entire populations using both conventional and molecular markers; a technique used successfully by Landau-Ellis et al (1991) to map the locus controlling supernodulation (nts) in soybean, and Nickell et al (in press) to map fap2, a gene conferring high levels of palmitate. Phenotypic ‘extremes’ can also be genotyped to associate one allele with one phenotype, thus suggesting the linkage of the marker with the gene (Michelmore et al, 1992).
A similar association of marker and gene can be determined by screening near-isogenic lines (NILs). The potential of NILs in integrating soybean conventional and molecular linkage maps was discussed in depth by Muehlbauer et al (1988). They calculated that if a BC5S1 NIL was screened with 100 randomly chosen loci, assuming polymorphisms existed between the recurrent parent and the donor parent at all loci, four loci should detect donor DNA and two or three of them could be expected to be genetically linked to the introgressed gene. Evaluating 63 NILs, each possessing an introgressed conventional gene with 12 isozyme loci, five presumptive linkages were observed by Muehlbauer et al (1989). Segregation analysis confirmed linkages of Enp (endopeptidase) with In (narrow leaflet) (9.38 +/- 1.55%) and Mpi (mannose-6-phosphate isomerase) with dt2 (indeterminate stem) (16.07 +/6.43%). Using 15 RFLP markers, Muehlbauer et al (1991) screened 116 NILs. Fifteen polymorphisms were observed where the NIL possessed the donor parent allele. Segregation analysis confirmed linkages of pK-3 with p1 (pubescent pubescence type) (16.2 +/- 5.0), pK-3 with r (brown seed) (14.3 +/−4.6) and pK-472 with Lf1 (5-foliate) (14.10 +/- 4.82).
A series of isolines containing one or two alleles for resistance to phytophthora root rot (Rps1-6) and the locus conferring ineffective nodulation (Rj2) were screened with 141 mapped RFLP markers (Diers et al, 1992b). At least one polymorphism was detected between each NIL and the recurrent parent. Segregation analysis of F2:F3 lines confirmed linkages of Rps1 to pK-418 and pA-280; Rps2 and Rj2 to pA-233 and pA-199; Rps3 to pA-186 and pR-45; Rps4 to pA-586; and Rps5 to pT-5. However, even though linkages have been established, the position of some of these markers relative to other mapped molecular markers remains ambiguous.
The map localization of qualitative genetic factors not only makes the map more interesting but provides the opportunity to speed the introgression of the characters from an exotic source into a cultivar or from one cultivar to another. The usefulness of the approach depends, of course, on the 'scorability' of the character and the tightness of the linkage between the marker and gene. One of the goals of this backcrossing program is to eliminate the donor genome and replace it with the recurrent parent genome. In practice, the amount of donor DNA flanking an introgressed gene is not reduced by backcrossing as quickly as statistically predicted (Young et al, 1988). The number of backcross generations required to achieve introgression and donor genome elimination can be dramatically reduced by selecting, not only for the markers tagging the gene of interest, but by selecting against the remaining donor genome markers (Beckmann and Soller, 1986). This cannot only reduce the time required to complete a backcrossing program but can result in a more pure recurrent genotype.
Mapping quantitative trait loci
A limited number of studies have been conducted towards identifying quantitative trait loci (QTLs) in the soybean genome (Table 1). Hard seededness, a quantitative trait that affects germination rate, viability, and quality of stored seeds, was evaluated...
