Association mapping (AM) or Genome wide association studies (GWAS), also known as linkage disequilibrium (LD) mapping, is a method used to identify marker alleles associated with phenotypic traits. It makes use of ancestral recombinations present in the gene pool of a species to identify marker/trait associations.
I would say you may need atleast 100 genotypes to do association mapping in self or cross pollinated crops. 150 and above would be good. Again the more the diverse genotypes and more number of markers in your set then better the mapping results would reflect. Other important informations that should be known are given below.
Advantages of Association mapping
Contrary to (Quantitative trait locus) QTL mapping, association mapping needs large numbers of molecular markers and uses recombination events present within the gene pool of an organism. Therefore, in association mapping, only those markers that are tightly linked to the locus of interest and that are within the extent of the Linkage disequilibrium (LD) decay of the genome will show significant associations. Thus the marker-trait associations identified through association mapping are very close or within the gene affecting the trait and results are applicable to wider germplasm.
Populations to be used
Various kinds of populations can be used for association mapping. Accessions collected from wild, gene bank collections, cultivars, and specialized populations [e.g., Nested association mapping (NAM), and Multiparent Advanced Generation Inter-Cross (MAGIC) lines] are commonly used in plants.
Importance of Linkage disequilibrium
For example, LD generally decays faster in cross-pollinated crops compared to self-pollinated, diverse populations compared to populations with narrow genetic background. Also, different genes and genomic regions in the same crop can exhibit different rates of LD decay. In the context of association mapping, the extent of LD decay defines the number of markers needed to conduct association mapping analysis and the expected resolution of identified associations. In self-pollinated crops where LD extends hundreds of kilobases, association mapping will have lower resolution, that is, a marker showing significant association could be far away from the actual mutation affecting the phenotype. In cross-pollinated crops where LD decays within few hundred base pairs, association mapping will have high resolution, that is, any marker showing significant association will be very close, or within the gene affecting the trait.
I think the previous answer is very informative and conclusive dear Khaled, however, I can say that the population size is very crucial in determining the resolution and density of the genetic map constructed, so as the size of the population is beyond 200 the map will be very dense.
Information seems to be generic. Although there are several mixed models are being built, in my opinion, association mapping in a highly self pollinated crop would mostly give a false positive associations.
For Genome wide association mapping in ryegrass (cross/open pollinated), how to phenotype them?
If I have 100 wild populations, leaf samples of 10 plants from each population are collected, then do I need to phenotype those selected 10 plants directly, or I need to make clones of each of those 10 plants for phenotyping?