I am new in microarray data analysis. Recently I wanted to use Limma to find different expressed genes between grade I and grade III, which you could see from the annotation file, but when I use the R code below to do this, I only got two different expressed genes. Could anybody tell me why? Is there anything wrong with my code? The data is available online at http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE2990.
and the annotation file is attached to this post.
rm(list = ls())
library(affy)
library(affycoretools)
library(simpleaffy)
library(gcrma, warn.conflicts = FALSE)
library(hgu133a.db)
library(a4)
datainfo
thank all of you very much, so you mean my result is right. I could noly got two different expressed genes when I compared these two groups?
Among those genes you can select to control the fdr at 5%, there seem to be 2 genes with an average |lfc|>2. That's it. I can't tell you if an fdr of 5% is reasonable or the focus on genes with strong average regulation (differential expression) is resonable. To my personal opinion, 5% fdr might be a little restrictive if the list is used to subsequently analyze properties of the entire list, and focussing on genes with |lfc|>2 is clearly too restrictive. If you release this criterion a little, say to |lfc|>1, you will get more genes in your list, that then includes genes that are not *so* strongly (but still considerably) regulated.
Thank you, jochen. So as for fdr, what value do you recommend? How about 10%?
It depends a little bit on the purpose of this list. For getting a hint of pathways and bilogical processes potentially playing a role in your experiment, 10% or even 20% might be ok. If the aim is to get a list that is subsequently elaborated candidate by candidate, 1% or 5% may be ok. IEventually, it is just your decision - there is no general "right" or "wrong".
- Badges
- Science topic
- Similar topics
- Biological Science
- Genetics
- Gene Expression