Quantitative trait normalization question

Hi y'all,

I'm in the field of population genetics and I'd like to perform
genetic association test.
Problem is I am more of a molecular person than a statistician.

I have bunch of phenotype information but I have very little knowledge
about how to process these info.

What I have learned so far is:

1. I have to normalize quantitative traits
2. I have to adjust phenotype of my interest to covariates

The only normalization technique I know and did was [log
transformation]
I've heard that [normal quantile transformation] is an effective way
to normalize quantitative trait values. But I do not even know what
this is, or how to do this.


What I did was:

1. log transformed quantitative trait values
2. adjusted for covariates using SAS/Analyst: ANOVA: Linear model,
placing the QT of my interest in [Dependent] box, covariates in
[Quantitative] box and sex, medication status, etc in [class] box.
3. Residuals were produced by step 2
4. Use residuals for further association test

What I'd like to know is:

1. What is the difference between putting sex, medication status,
drinking, smoking in [class] box and putting those in [group] box?
What is the right way to adjusting for those values?
2. How does 'normal quantile transformation' works and how do I do it
using SAS or any other package?

Can anyone help me on this please?


Charles

.