[BioC] Variance correction, coxph or cph model

[aedin]

Hi Bioc
I am using cph in the library Design so that I can compare the log 
likehood ratio test for different nested cox models.  I am testing about 
50 genes. Whilst I can use p.adjust to correct for multiple testing, I  
get a few gene with low expression levels, which may have low variance 
by chance. Is there an ebayes (limma) like correction I can apply?     
Or can I modify this for cph?
Thanks
Aedin

-- 
Aedín Culhane
Harvard School of Public Health, Dana-Farber Cancer Institute