what is inverse normalized RNA-seq expression and how to do that?

Question

[piyushjo]

Hi,

I just found out this new term "inverse normalized gene expression". I am reading a paper where they use inverse normalized expression instead of rpkm or vst normalized data for cox regression. Below is the statement from the publication.

It is common to find extreme outliers in large RNA-seq datasets, which can negatively impact survival regression analysis. In order to correct for these outliers, the expression values for each lncRNA were inverse normal transformed, a procedure that increases the sensitivity and specificity of regression analysis using RNA-seq expression values [59].

I also went to through their citation paper, but couldn't find a program to do it. Didn't understand the mathematics.

Can anyone help me transform my rna-seq count or vst normalized data to inverse normalized? Also how can I do that to micro-array data?

Thanks. P.S.: Updated with information.

[WouterDeCoster]

I am reading a paper where they use inverse normalized expression instead of rpkm or vst normalized data for cox regression.

I think it would be useful if you would provide a link to that paper.

[Kevin Blighe]

Here is a previous thread: Inverse Normal Transformation Of A Quantitative Trait In Gwas

[piyushjo]

Hi Kevin. Thanks for your response. I did see that particular thread and the cited paper. However, I didn't understand how to apply the suggested formula to RNA-expression.

qnorm((rank(x,na.last="keep")-0.5)/sum(!is.na(x)))

For example, what is X? Gene expression value or gene name, and what would be the output of it? Also if there is an R package that will transform my micro array data or I would need to use this formula for each sample?