I'm interested in checking if some genes I have are tissue-specific in their expressions. I'm working with TPM normalized counts and want to use the Tau statistic to measure specificity. I think this article has the most comprehensive description of Tau. In the same article they work with RPKM normalized counts which then they log-transform before proceeding to Tau calculation. In my case I also log-transform TPMs, but also apply Quantile Normalization before carrying out some other downstream analyses. My question is if I need to go back to the expression matrix I had before Quantile Normalization or if it should be fine to calculate Tau even after Quantile Normalization.

Is it the same Tau as in Kendall's tau correlation coefficient ? If it is so, my understanding is that

Kendall's Tauis calculated based on ranks, which are preserved after quantile normalization. So it should lead to identical results to calculate Tau before or after quantile normalization.The URL of the article seems to have expired. Would you mind posting the current address?