Hi Tilestat. Thank a lot for your videos. I'm trying to apply the Fisher's Exact test on RNAseq count (using EdgeR exactTest function ) but so far I can't understand how the p-value is calculated. I'm not able to apply your example on a gene count matrix (I'm feel so stupid 😭😭😭...) Can you give me some suggestion?? In any case thank a lot for all your video, I appreciate so much... thanks again
Sorry, I have never used this type of test on RNAseq data. However, I have previously used Poisson regression (or NB regression) on such data that I have a number of videos on.
Maybe if you "translate" read count into a categorical variable? Say you have tissue 1 and tissue 2, and you want to see whether expression of your gene is associated with tissue type. You could define a threshold of counts, at which you consider the gene upregulated. Then you can do the matrix like in the video (smoker, non-smoker = tissue 1, tissue 2 and cancer, healthy = upregulated, not changed). I would say it depends on your research question.
![Noob To Max With DRAGON REWORK In Blox Fruits [FULL MOVIE]](http://i.ytimg.com/vi/LnBMOinoOvA/mqdefault.jpg)
Really having a great benefit from following your channel!
Great to hear that.
Very good explanation, thank you
Hi Tilestat.
Thank a lot for your videos. I'm trying to apply the Fisher's Exact test on RNAseq count (using EdgeR exactTest function ) but so far I can't understand how the p-value is calculated.
I'm not able to apply your example on a gene count matrix (I'm feel so stupid 😭😭😭...)
Can you give me some suggestion??
In any case thank a lot for all your video, I appreciate so much...
thanks again
Sorry, I have never used this type of test on RNAseq data. However, I have previously used Poisson regression (or NB regression) on such data that I have a number of videos on.
Maybe if you "translate" read count into a categorical variable? Say you have tissue 1 and tissue 2, and you want to see whether expression of your gene is associated with tissue type. You could define a threshold of counts, at which you consider the gene upregulated. Then you can do the matrix like in the video (smoker, non-smoker = tissue 1, tissue 2 and cancer, healthy = upregulated, not changed). I would say it depends on your research question.
I thought that we only use FET if the expected values is less than 5 in every cell?
For a Chi-square test, the expected values should be 5 or greater in at least 80% of cells