SNP data analysis

Question

[caspase8mach]

Hi,

I would like to analyze SNP data for close to 100 samples. Each sample has close to 100 SNPs, and each SNP is related to a trait.

How can I perform an analysis to come up with patterns, co-occurrence of traits, etc?

If possible, I would be thankful if some R based solution is suggested.

Thanks

[WouterDeCoster]

How can I perform an analysis to come up with patterns, co-occurrence of traits, etc?

It's unclear what you aim to achieve, please elaborate. Aspecific question like this rarely get answered accurately.

[LauferVA]

I second this. If you are uncomfortable with the statistics, that's totally OK. Rely on your understand of the biology to help guide us to the kind of test you want to design.

Example: There is a paper (Corradin et al 2014) describing something called the multiple enhancer variant hypothesis. Essentially the idea is that multiple variants all acting on nearby enhancer elements that affect expression of just one gene confer much greater risk of autoimmunity than single variants in such a region.

If this were your question, you should describe that hypothesis as clearly as you can. Then we could help you design a statistical test (or machine learning approach) that would work to optimally discover that kind of pattern, if it is indeed found in the data.

Without such information, we might recommend (in error) that you conduct single variant association tests (test each variant for significance independently of the others).

[caspase8mach]

All right, let me re-phrase my question:

I have performed SNP analysis on 100 samples and have attributed the outcome based on the genotype to specific traits, for example, height, hair length, etc etc. So, for each sample, there are multiple outcomes (generally 3 - positively affected, not affected, negatively affected). Can I employ statistics to see if there are some traits which co-occur? Is there some more information I can get out of such data?

Thanks