Alicia Schep

I unfortunately still can't reproduce that output or the failing test with those package versions, making getting to the bottom of this issue difficult. Do any of the other tests...

Thanks, it looks like those are the same tests from tests/test_var.py run earlier. Seem like the problem is likely with the calculateCov function in the nucleoatac/multinomial_cov.pyx file, although I am...

For troubleshooting, perhaps figuring out more specifically why the tests are failing will help. If you try the following code within python in the main NucleoATAC directory (or edit file...

Ok that helps confirm that the problem is with calculateCov function specifically. After running the first two blocks of code above (through signaldist = ...) can you try: ```python import...

Yes that is the expected value. 'signaldist.analStd()' should just do those two computations and then take the square root of var -- very weird that it seems to work when...

I am not sure what that ifelse would test for, as i am not sure whether in other cases the function could give an other type of spurious signal. I...

One possible fix is using a docker container... I started looking into how to make one, and found that it seems like someone else has already created one: https://biocontainers.pro/registry/#/showImages?domain=quay&repository=nucleoatac&starred=false&modified=18%2F01%2F2017

Thanks @vatese for figuring this out! I guess I'm still not sure how to actually make the code compatible with the earlier version of gcc, but at least now we...

@sameet Unfortunately I'm not sure what the best next step would be :slightly_frowning_face: I still don't know the underlying cause of this bug, other than that the gcc version seems...

Are any figures generated from Step 1? If so, if you can post them, those can be useful diagnostic (sometimes there are errors because fragment size is really different from...