I've been using the single-condition ASE_detection and it seems to be working fine but I'm using a higher number of permutations, which leads to some of the jobs timing out (expected, I will re-run for longer later). Those genes output an empty file with just the column names "gene" and "p-value" and no numbers. However, some of the output files have gene name, what i assume is the likelihood statistic? (mentioned in another issue), and a p-value of zero. Histograms of the distribution of the unlabeled variable in the output file are only positive for genes with a p-value=0, and genes with a non-zero p-value seem to have largely negative values for that variable. Any idea what might be going on?
Also the paper suggests additional covariates could be incorporated into the model but I don't see an obvious way to add them and would appreciate any suggestion! Would consider adding age, sex, and ancestry PCs if possible.
Hello,
I've been using the single-condition ASE_detection and it seems to be working fine but I'm using a higher number of permutations, which leads to some of the jobs timing out (expected, I will re-run for longer later). Those genes output an empty file with just the column names "gene" and "p-value" and no numbers. However, some of the output files have gene name, what i assume is the likelihood statistic? (mentioned in another issue), and a p-value of zero. Histograms of the distribution of the unlabeled variable in the output file are only positive for genes with a p-value=0, and genes with a non-zero p-value seem to have largely negative values for that variable. Any idea what might be going on?
Also the paper suggests additional covariates could be incorporated into the model but I don't see an obvious way to add them and would appreciate any suggestion! Would consider adding age, sex, and ancestry PCs if possible.
Thank you!
Chani