This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Correlation analysis for low expressed genes

Dear all, I am doing a correlation analysis on few targeted genes and I found there some genes that are low expressed. For example, a gene expression data for two gene with 100 subjects below:

        Gene A   Gene B 
Subject1  0       45
Subject2  0       42
Subject3  0       55
Subject4  0.1     78
Subject5  0.03    66
   ...
Subject98 0.09    69
Subject99 0       47
Subject100 0      44

However, there are many zero values in Gene A, and only three subjects are detected little gene expression on Gene A. Here is my question. Is it reasonable to use Spearman correlation to capture the relationship between Gene A and Gene B? I would greatly appreciate it if you kindly give me some feedback.

correlation analysis
• 1,110 views • link
written by kmsh410 • kmsh410

1 answer

One option is to use the samples with only non-zero expression values. Other option is to perform permutations and calculate a p-value.

• 0 views • link
written by GouthamAtla GouthamAtla

Appreciate your suggestions. I will take a look some permutation approaches and other tools that could analyze the zero-inflated data.

• 0 views • link
written by kmsh410 • kmsh410

Log in to answer this question.

More posts like this