Hello Biostars! I am happy to be writing my first post after several years of anonymous browsing and can't be more thankful for all the help I have received so far. I would really just like to hear your opinions on the design of my experiment.
I have currently collected about 20 human tissue samples from patients suffering from chronic bacterial infections. Along with each sample, there are about 15 different metadata variables. I would like to use RNA-seq to look into the transcriptome of these bacteria and which genes/networks are important in the establishment of chronic infections. For example, in these samples, there are 5 or so that still didn't heal after 12 weeks of treatment.
From what I have read, it seems that DEseq and similar approaches are specifically used to tell the difference in gene expression between two very clear conditions.
So, my question is...is there some way to include all these different metadata catagories into the analysis to get a picture of which metadata catagories play a bigger role. Something like some type of coexpression network that could take into account all metadata and try to figure out which genes change across a category of metadata?
I know this is very broad and I'm not looking for long specific descriptions, I am just interested to hear your opinions.
Thank you :)
I heard about WGCNA by which you can infer the most significant genes within each datasets. this R package offers several options that I think make sense in your case.
Hi Fereshteh! Thanks for the suggestion. That package does look like it would be useful :)