!. Translocation Mapping:
m: the recessive mutation. Backcross the gametes produced from this individual to either parent. When a gamete carrying the two wild type chromosomes is backcrossed to the translocation homozygotes, the F1 is a normal translocation heterozygote. All fully fertile (non translocation heterozygote) plants are translocation homozygotes, and are fully fertile. The translocation heterozygote (M/m) plants have the translocation in repulsion with the mutation. The plant is semisterile (why?). The seeds that produce mutant plants (m/m) all produce fully fertile progeny when backcrossed to the mutant, non-translocation-carrying homozygote. The way you link mutations to translocations is through their repulsion segregation in crosses back to translocation homozygote parents and the homozygous recessive tester stocks.
The problem: this sort of mapping only works when the mutation is within ~15cM of the translocation, unless you're really good at making lots of crosses. That would require around a dozen uniformly recombinationally spaced translocations per chromosome.
Bob Eslick, my predecessor, thought to improve the efficiency of this process by coupling recessive lethal mutations to translocation breakpoints. This coupling of lethal mutation to translocation required maintenance of the translocations as heterozygotes, and then crossing to them with mutant stocks. In the progeny, the ltTr/leTr individuals died, the translocation heterozygotes had the dominant phenotype and were semisterile, and the m/m individuals were fully fertile. In all crosses where the gene was unlinked to the lethal translocation, the segregation of the mutant phenotype followed normal 3/1 expectations. In crosses where the gene was close to the translocation/lethal mutation complex the expected ration would be 2/1 (translocation homozygotes died, because of their linkage with the lethal recessive).
In this case as well, coverage of the genome was a problem. Also, think of the number of crosses you would have to make to distinguish between a 2:1 vs a 3:1 ratio. People used to work really hard to map a gene.
Trisomic mapping
Primary trisomics can be prepared by crossing a tetraploid to a diploid, producing a triploid. The triploid is then self-pollinated or pollinated by the diploid parent. Since aneuploid gametes are strongly (but not completely) selected against, most of the fertile seed will be either diploid (2n) or trisomic (2n+1). Do roottip squashes of each plant to identify the trisomics.
Take a homozygous recessive mutant. Cross it to (in the case of barley) each of the seven primary trisomics. From each cross, select trisomic progeny. For 6 of the seven chromosomes, the mutant gene will be heterozygous (Mm). For the cross in which the gene lies of the triplo chromosome, the mutant gene will be MMm. Pollinate each of these seven cases with mm pollen. Pencil out your expectations for diploid and trisomic progeny.
Problem with this approach: demands enormous numbers of good roottip squashes and chromosome counts. Also, this only gets your gene mapped to a chromosome.
Addition/Substitution Line Mapping
To this point, we've studied the use of allelic variation to identify and map gene action. This is almost always difficult- alleles are often only subtly different. With a reasonable frequency it's impossible. No allelic variation is available for as many as 30% of barley's genes. The problem of allele mapping is even greater in bread wheat, a species with far less intrinsic variation than barley.
If a portion of a chromosome was placed in an alien nucleus, we could identify the genes on that portion of that chromosome by their unique structural characteristics relative to their homoeologous relatives. Ernie Sears recognized this fundamental truth in the late 1940s. By the 1950s he'd begun development of a complete series of wheat-rye substitution lines. He found a wheat genotype, Chinese spring, to be receptive to genomic manipulation. Chinese spring was crossed to several rye genotypes, and wheat/rye substitution lines were produced. Once characterized, these enabled researchers to identify the rye chromosome that carried a specific rye gene.
Deletion Line Mapping
The process of embedding partial genomes of grasses into wheat became common in the 1960s and 1970s. Some of this work was done for simple, practical gene transfer purposes. Our best wheat streak mosaic virus resistance gene derives from one of these programs. Cytoplasm substitution has also been frequently done using interspecific crosses, in the hope of discovering either a better plastid genotype or a more useful cytoplasmic male sterility system. In the process of generating one of these cytoplasm substitution stocks between Aegilops cylindrica and wheat, Endo noticed enormous levels of male sterility. Endo and Gill attempted to produce substitution lines from crosses between Aegilops cylindrica and wheat. A gene on cylindrica's chromosome 2 partially suppresses induction of male sterility. Substitution lines carrying other cylindrica chromosomes generate high levels of sterility and chromosome-level mutation. Here's Gill's most recent discussion of what happened.
Genes from sharonensis, speltoides of cylindrica all can induce chromosome fragmentation leading to deletions and partial to complete male sterility. The cylindrica effect is more mild than that of the other two species. The breakage-bridge-fusion cycle of chromosome breakage appears to be involved, we'll talk more about this in the McClintock lectures.
The practical outcome of this was the development of a whole new type of genetic mapping tool for wheat. Nested deletion stocks were prepared for each chromosome, after backcrossing out the cylindrica chromosome and fixing different deletions in different lines. These are now the primary mapping tools for the wheat genome mapping project. Several laboratories from around the country are collaborating to place several thousand EST-based RFLPs on the wheat linkage map. Some of the results can be viewed in recent manuscripts by Bikram Gill and Kulvinder Gill.
Given that we have good maps for most of the crops, it
might be reasonable to ask whether we need to resynthesize a map every
time we decide to map a gene of interest. It is not absolutely necessary.
Consider a gene with two alleles, one with a positive
impact on plant height and one negative, in a population of F2 plants.
If the gene shows dominant gene action and is the sole factor segregating
in the population which impacts the trait, then 3/4 of the plants will
be tall, 1/4 short. If we perform single seed descent on the population,
then eventually the plants will show a bimodal distribution for plant height.
All the tall ones will be AA, all the short ones aa. If we then bulk
DNA from the talls, and bulk the DNAs from the shorts, we should have a
uniform distribution of allelic states for all markers unlinked to the
plant height gene within each bulk. However, the closer we come to
the gene, the more the allelic distribution within the 'tall' bulk should
resemble the tall parent, and the more the allelic distribution for the
short bulk should resemble the short parent. If you know the chromosomal
location of the markers you're using, you then know the chromosomal location
of the gene. This is called 'Bulked Segregant Analysis' (BSA).
Consider a trait controlled by two genes of similar magnitude.
Will Bulked Segregant Analysis work? What sorts of trait heritability
do you need for BSA to be effective? What sort of mapping technology
works best for BSA?
Tailed distribution analysis provides an alternative to
BSA. This approach demands that you take the lines representing the
tails of the phenotypic distribution and run them individually against
a set of mapped, informative markers. You can then test the frequency
of alleles within each population and identify skewed segregation.
What marker limitations exist for this design?
If multiple genes modify the trait of interest, neither
approach will work well. Are there ways to think about this problem
which will make the approach work even if several (say 4) genes modify
the trait in your target population?
These approaches suggest that you know nothing about the physiological basis for the trait. A more intellectually appealing approach involves the use of
Candidate Genes
Candidate genes are those genes which might reasonably
be expected to affect the expression of the trait you measure. We
enjoy a very limited view of the physiological basis for the traits we
study. The geneticist, although happy to characterize the action
of a useful gene (through QTL analysis) and denote its chromosomal location,
would be happier if he could say 'the short allele is due to a mutation
in a specific translation factor'. Thusfar we're still pretty far
away from this, but we're getting closer.
Expressed Sequence Tags (Ests) coupled with SNP analysis
provide a potentially useful way to identify candidate genes. Take a look at the
Graingenes
wheat or
barley EST program
and think about how you would couple QTL analysis with gene expression analysis
and mapped ESTs to begin a search for candidate genes key to a phenotype
interesting to you.