Entering edit mode
3.9 years ago
qstefano
▴
20
Hello there,
I am looking for a statistical test to be able to make inferences on percentile values retrieved from gene expression in RNA-seq.
I have several genes with different percentile values (conditions), i need a test where the null hypothesys is that these values are not statistically different.
Is there a specific test for percentile values? looking around I found nothing.
Thanks.
Late to the question and not sure I understand what you're trying to do but in case this helps someone, notice that percentiles can be converted to z-scores if we're dealing with normal distributions. Otherwise, it seems you may want to compare distributions, in which case you can use the Kolmogorov-Smirnov test.
It looks like you are trying to do differential gene analysis with normalized gene counts. A quick and dirty approach would be t-test if it is normally distributed, the decent way however, is to obtain the original count and start there. hth