Plant & Animal Genome VI Conference

S12

CANDIDATE GENES AS QTL

Department of Genetics, The Silberman Life Sciences Institute, The Hebrew University of Jerusalem, 91904 Jerusalem, Israel

The candidate gene hypothesis proposes that a major portion of trait quantitative genetic variation is caused by functional variation in the genes directly involved in trait development or physiology. The hypothesis is supported by the incompatibility of widespread "random" pleiotropy with evolvability; by a high quantitative mutation rate; by innumerable positional cloning and knockout studies showing that mutant trait and disease phenotypes are very generally caused by mutation in candidate genes; and above all by the unexpectedly large (although still absolutely small) number of studies showing candidate gene association with quantitative genetic variation. Even if most quantitative genetic variation is due to candidate genes, only a small proportion of plausible candidate genes will be associated with quantitative effects in any particular agricultural population. This is due to the small number of QTL actually contributing significant genetic variation in any population. Thus, many candidate genes for a particular trait must be screened to identify a few QTL; and different candidate genes may be serving as QTL in different populations. It is justifiable, however, to start today with the available candidate genes, and add to these as new candidates are uncovered. Searching for candidate gene effects requires large scale identification of genomic sequences and polymorphic sites within known candidate genes, and definition of intragenic haplotypes for maximum power. The latter is facilitated by the small number of alleles at any given locus in a typical agricultural population, dictated by the small effective population size. Resource population design and statistical analyses for optimal candidate gene analysis also differ from those required for linkage mapping of QTL. Thus, implementation of a serious candidate gene approach, gets an assist from population genetics, but still requires major infrastructure development by the genome mapping community. A candidate gene approach is compatible with small experimental or commercial populations, and leads directly and naturally to MAS. In this way, it supplements the limitations of the more general linkage mapping approach, and provides a promising alternative route for genomic applications in plant and animal populations.