SNPs screening (Associations between each SNP and outcomes)

This module tests for association between each SNP and each phenotype.

SNP data could be from a .ped file. For each SNP, three models: additive, dominant and recessive will be tested.  If the phenotype is a continuous variable, this module will also generate adjusted mean (and standard error) of phenotype for each genotype (AA, AB or BB).

The output table includes following columns:

         SNP name,

         Phenotype name,  

         Group (if a stratified variable was specified)

         N0, N1, N2:  number of observations used that had genotype=AA, AB, BB respectively  

         (add), se(add), p(add):  regression coefficient, standard error and P-value from additive model  

         (dom), se(dom), p(dom): regression coefficient, standard error and P-value from dominant model  

         (rec), se(rec), p(rec):  regression coefficient, standard error and P-value from recessive model  

         lsmean0, lsmean1, lsmean2: adjusted mean for people with AA, AB, BB genotype respectively  

         fdrcut:  FDR cut point for P-value from additive model (Benjamini and Hochberg 1995

:  sort p values [k=1 to m], start at biggest p value, stop when p <= 0.05*k/m)

What its used for

This module was used to do an initial screening over a large amount of SNPs. It helps us to quickly identify most significant SNPs.

A screen shot of sample design:

t9_input.gif

A sample output table:

t9_snp.gif