Basic Conception On Gwas

Question

[Lds]

Hi all,

I'm a new on Genome-wide association study. I just want to know some simple ideas on GWAS as follows,

1. What's the statistical power in GWAS, how to calculate it?

2. Some studies may contain multi-stage GWA-studies, why should they do this?

3. For the Manhattan plot, the values for Y-axis may be Fst or -log(P), I know how to calculate the Fst, but for -log(P), what P stands for? How to calculate it, I mean the math approach.

4. There may be a regression model for the outcome, such as linear regression for a quantitative trait or logistic regression for a qualitative trait, what the model looks like?

Maybe some can show some reference here to guide me to understand such questions filled in my head. Thanks in advance.

Linus

[Zev.Kronenberg]

I would suggest taking a look at this lecture:

http://www.stats.ox.ac.uk/~mcvean/gwa4.pdf

it is very informative.

[Zev.Kronenberg]

I would suggest starting with a good review article.

http://www.nature.com/nrg/journal/v6/n2/abs/nrg1521.html

Genome-wide association studies for common diseases and complex traits

Answer 1

[Darren J. Fitzpatrick]

STATISTICAL GENETICS: Gene Mapping Through Linkage and Association is an excellent introduction.

It has chapters addressing your specific questions!