Hi
I use limma to find DEGs between chemo-resistance and chemo-sensitive samples, in limma output I found that e.g one of my genes has log fold change= - 0.87, Now I have a question, Does this down regulation happened in my chemo-resistance or chemo-sensitive samples?
It worth to notice that I don't find any relation between this gene and my hypothesis in texts and papers.
I apologize for my minor question
Since you've setup your design with an intercept, the intercept becomes the first level of your "chemosensitivity" factor, which is "Rx Insensitive".
The second coefficient tests the log fold change of "Rx Sensitive" over the intercept, so your reported logFC's correspond to "Rx Sensitive - Rx Insensitive"
To confirm:
Thank you very much
colnames(design)
[1] "(Intercept)" "factor(sensitivity$chemosensitivity)Rx Sensitive"
According on your explanation the log fold change correspond to Rx Sensitive ,In my case log fold change= - 0.87 this negative log fold change is related to sensitive samples, It means that in this test, Rx Insensitive imply control and sensitive is case
please correct me if I'm wrong
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Please read http://www.bioconductor.org/help/support/posting-guide/. In particular, post some code.
It all depends on how you setup your model, which (as Aaron has pointed out) you have not shown us.
Also, it always helps to plot your data. So, you have identified the gene of interest, now you can plot its expression across your groups (sensitive vs resistance) -- the direction of the fold change should be clear, then.
Thank you for your reply design <- model.matrix(~factor(data$chemosensitivity)) factor(data$chemosensitivity) [1] Rx Sensitive Rx Sensitive Rx Insensitive Rx Insensitive Rx Insensitive ... Levels: Rx Insensitive Rx Sensitive fit <- lmFit(data, design) keep <- fit$Amean > median(fit$Amean) ebayes <- eBayes(fit[keep,], robust=TRUE, trend=TRUE) tab <- topTable(ebayes, coef=2, adjust="BH",n=100)