RNA Seq- Performing pearson correlation with clinical variables

Question

[shreev10]

Hi All,

I am Ashok. I have a question with regards to performing a correlation between RNA seq data and clinical variables. We have a total of 40 samples with well matched clinical annotation. We performed RNA seq and we are trying to correlate the expression data from RNA seq with clinical variables such as Blood pressure, Tumor stage and grade. My questions are:

1. Is it right to perform these type of correlation? My clinical variable data is normally distributed but I am aware the sequencing data is based on different distribution.

2. If we can perform this, what type of data should I use? Raw reads or normalized counts?

I would be thankful if you can help me with these questions.

Thank you. Ashok

Answer 1

[ewre]

Maybe spearman correlation between gene expression FPKM value and your clinical variable is a good choice. Also, if you find both gene A and gene B is correlated with, say, blood pressure. You'd better check whether gene A and gene B are correlated or not.