Using Expression of a Gene as a Linear Regression Variable

Question

[jamie3355]

Hi there,

I am very new to bioinformatics, so apologies in advance for any lack of clarity in my question.

I am attempting to investigate genes that may influence the expression of CCR9, in an RNAseq dataset obtained from a large cohort of children newly diagnosed with T-ALL.

The data has been combined into a unified count matrix using RSEM and I plan on using RStudio to process these data.

My initial thoughts were to create a condition within this group based on high and low expression of CCR9 and carry out gene expression analysis using DESeq2. I am now wondering whether it may be better to use another package, where I could use CCR9 expression as a regression analysis variable? I was wondering whether Limma would be a suitable tool for this use case.

Many thanks in advance for taking the time to read my question and thank you in advance for any suggestions and advice you can provide.

Answer 1

[dsull]

DESeq2 can do regression with continuous variables.

Another simpler option if you have a large cohort is simply to do the pearson or spearman correlation between CCR9 and every other gene.

[i.sudbery]

I don't know why, but people often say that limma-voom is more suitable for continuous variables than DESeq2. Not sure of the reason.