RNAseq and future analysis

Question

[Roy]

Hello all,

I am new in RNA analysis and am trying to measure the expression levels of a patient given the RNAseq BAM file. I used featureCounts to get the gene level counts for this sample, but now I'm kind of stuck on my next analysis (DESEQ2, cibersortx?). I am wondering if I should also get some control data from the same tissue and run DESEQ2? So I would have my one patient data and ~5 controls then run the expression analysis...is there a better method or direction someone could point me to? End goal is to observe whether key genes are highly expressed or not.

Thanks, Roy

[ATpoint]

What do you expect to get out of a single sample? What is a "key gene"? Expression levels in RNA-seq do not really reflect biological relevance.

[Roy]

Key genes refer to a set of genes that are of interest to me since they can be targeted for cancer therapy. There have been multiple tests of all sorts done for this individual, so I am analyzing the raw data for any signs of potential therapeutic markers that may have been missed. (ex. if EGFR is highly expressed, theres a drug for that)

[K.patel5]

Differential expression analysis followed by some form of functional analysis (GSEA or overrepresentation analysis) on the significantly differentially expressed genes (low p values) with a background set of genes (all genes found in RNAseq) is a fairly standard and ~non-bias way of identifying candidate genes. Don't think you have enough samples for anything too exotic.

Answer 1

[swbarnes2]

RNASeq is strongly affected by batch effect. Because of this, you can't just grab a control somewhere and compare; most if the changes you see will be batch effect, not biological. If you want to do DE analysis, you have to plan your experiment to have all the replicates and controls you need from the start.