Illumina Seminar on developing MAS on agrigenomics in plant
I just came back from a seminar on plant Marker Assisted Breeding organized by Illumina yesterday. Since I'm waiting for a 8-hours script to complete (blame my bad programming skills), I will post something about the seminar. Last month, Illumina has announced to give away 10G of sequences to any Malaysian scientist who come up with the best 5000-words proposal. The deadline is 31 Aug 2010.
Back to the topic. The speaker is Dr. Richard Hodgson from Illumina US. He has vast experience in developing disease diagnostic methods for agriculture and aquaculture. He got involved in many breeding projects such as chili, coconut, shrimps and now he has a liking in durian. He introduced a relatively new approach called Genomic Selection(GS).
Unlike Marker Assisted Selection(MAS), GS is based solely on genotyping and estimation of breeding values. It has the advantage of capturing small gene effects not detected by QTL mapping. First, the breeders must have a large training population with known genotypes and phenotypes. It takes advantage of the cheap genotyping to screen tens to hundreds of thousands SNPs markers for large number of seeds/seedling. The number of markers depends on the diversity of the populations. The more diverse, the more markers needed. The selection is based on how closely the genotype matches the training population and breeding values is estimated based on predicted phenotype. Without phenotyping, the breeding and selection cycle is reduced significantly.
Here is a good introductory article about GS here. "In simulations, the correlation between the true breeding value of unphenotyped experimental lines and that predicted by genomic selection has reached 0.85. Genomic selection accuracies depend on a trait’s underlying genetic architecture, the level of linkage disequilibrium in the crop population relative to the marker density available, and the statistical methods used." Another paper mentioned why GS is better than association mapping.
Studies in maize and wheat has demonstrated success using GS. The question is how well does it work? How to apply it in non-model plants? Does it require a linkage map and location of each markers on the chromosomes? One thing for sure, you need to assemble a consortium, sequence a lot of varieties/lines and design a SNP chip for this purpose. Here's where Illumina play an important role in providing technologies bla bla bla... zzz.
Posts
