[BioC] Finding genes with significant linear regression, or trend

Assaf Oron aoron at fhcrc.org
Tue May 20 23:03:51 CEST 2008

Daniel hi,

You can also try the function "lmPerGene" from the new package GSEAlm. 
This function allows for continuous covariates without recoding them as 
factors. The function returns an object, say "mylm".
To get the t-statistics, divide "mylm$coefficients" by 
"sqrt(mylm$coef.var)". Note that the coefficients and their variances 
are matrices with the first row being the intercept. You can then take 
these t-statistics to your favorite significance/FDR sofware. You can 
also use it for gene-set-enrichment analysis (see the GSEAlm vignette 
for more details).

If you try to use the function and have more questions, you can write me 


More information about the Bioconductor mailing list