Input data format

[Lila14]

I was using the test data and noticed that the vast majority of values in the signal files are either 2 or 16. Was there some pre-processing done to this data?

[guanjue]

Yes, the signal was the -log10(p-value) based on the negative binomial model after normalized by s3norm. Since it is -log10(p-value), the 2 and 16 represent the p-value = 0.01 and p-value=1e-16.

[Lila14]

Why are these the only p-values?

…

On Wed, Oct 17, 2018 at 11:52 AM Guanjue Xiang ***@***.***> wrote: Yes, the signal was the -log10(p-value) based on the negative binomial model after normalized by s3norm. Since it is -log10(p-value), the 2 and 16 represent the p-value = 0.01 and p-value=1e-16. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#7 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ANh0A6Yuceus0Ckzogwt0ZHom_WUFHlfks5ul1I4gaJpZM4XkBsv> .

[guanjue]

In the test data, we have input data for our sample, so we used p-value to take into account the local background changes of the input signal.