I'm working on a population genetics of crane using 10 polymorphic microsatellite markers from cross-species amplification. I was using FSTAT for LD analysis. When I analyzed LD within population (one from 11 individuals and another one from 17 individuals), I didn't found LD but....
when I analyzed them together, I found 1 pairwise (from 45 pairwise) that exhibits a significance of LD.
Could you explain me why 1 pairwise was found?
According to the significance of LD, I want to know which locus I should select for genetic diversity analysis or 2 loci should be exclude?
Sorry for my poor English and thank you in advance for your suggestion.
Rangsinee
Hi Ransgsinee,
In most cases, researchers want to test for LD between markers so they can pick a set of loci that are effectively independent of one another (i.e., that are unlinked), as this is an assumption of some population genetic analyses.
The correct way to do this test is within a single population--ideally in a geographic area that is equal or smaller than the panmictic unit. You did this in both populations separately, and did not detect LD, which suggests that your 10 markers are located on different chromosomes, or at least in regions of the same chromosome that recombine fairly often. So I don't think that you should exclude any of these markers in further analyses.
Population structure probably explains why you found LD in the sample when you combined the geographically separated populations (i.e., 2-locus Whalund effect). Consider two loci (1 and 2) each with 2 alleles (a and b). in the first population, genotypes 1a and 2a are most common, while in the second population, alleles 1b and 2b are most common. So if we just pool all of these individuals and calculate LD, we see a non-random association of alleles among the loci.
Hi Ransgsinee,
In most cases, researchers want to test for LD between markers so they can pick a set of loci that are effectively independent of one another (i.e., that are unlinked), as this is an assumption of some population genetic analyses.
The correct way to do this test is within a single population--ideally in a geographic area that is equal or smaller than the panmictic unit. You did this in both populations separately, and did not detect LD, which suggests that your 10 markers are located on different chromosomes, or at least in regions of the same chromosome that recombine fairly often. So I don't think that you should exclude any of these markers in further analyses.
Population structure probably explains why you found LD in the sample when you combined the geographically separated populations (i.e., 2-locus Whalund effect). Consider two loci (1 and 2) each with 2 alleles (a and b). in the first population, genotypes 1a and 2a are most common, while in the second population, alleles 1b and 2b are most common. So if we just pool all of these individuals and calculate LD, we see a non-random association of alleles among the loci.
Hi Rangsinee,
I think this discussion linked below may help to address your question. Good luck
Best
Ali
https://www.researchgate.net/post/What_is_the_relationship_between_linkage_disequilibrium_and_population_structure
You probably need bigger sample sizes to test for LD, like 30. As correctly noted by Sean, do not mix data from differentiated populations. If two loci demonstrate significant LD, after correction for multiply testing, you need to exclude only one of them.
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology