I know that when two variants are close together and you've done an association study, both can come up as showing evidence for association just because they are in close proximity. I read that conditional association analysis can account for the effect of one variant on the other.

How is conditional analysis done, and what value tells you whether or not the variants are independent of one another? Does the odds ratio tell you? If so, what value should it be?

you may want to be a bit more clear by giving an example for contextual purposes.

Conditional analysis can be done using several different publicly available software packages. For instance, plink has the capability to do conditional analysis as does GCTA.

 

You look at the p-value of the associated snp after conditioning for the lead snp and see if its still has a significant association. If it does, then it has some independent effect on the trait.

you may want to be a bit more clear by giving an example for contextual purposes.

 

Conditional analysis tests snps that are in linkage disequilibrium with a snp that is associated with a trait to see if that the other snps are independently associated with the trait or simply associated as a result of being linked. Coming up with an example in this context would be slightly difficult, especially if they are not familiar with the concept.