leopoldc / boost-hic Goto Github PK

As tested in different size of input matrix, profiling the running time of function fastFloyd,
I found that its time complexity is extremely high at about O(n^5).

Here is the test result

So if I want to boost the Hi-C data in chromosome 1, It nearly costs 80 hours.

Is there anyway to accelerate this function, parallelizing or writing in C?

The README says that it uses python2.7, but the scripts seem to actually be written for python3. Additionally, I found some lines in HiCutils.py have tab indentation while others have space indentation. This caused it to break with some python implementations.

Hi @LeopoldC. Thank you for sharing this tool. I'd like to ask if there's a way to convert the h5 output matrix to hic or cool format. I've so far tried hicConvertFormat, but it's not working.

Thank you.

Hi @LeopoldC. While running Boost-HiC, the following warning is encountered, although the job is still running. I'm wondering if this is something one has to worry about, since the analysis has been running for almost 2 days with only filteredbin.txt, inputmat_filtered.hdf5 and inputmat.hdf5 in the output directory.

main.py:26: RuntimeWarning: divide by zero encountered in true_divide
FFmat=np.power(HiCutils.fastFloyd(1/np.power(normmat.copy(),alpha)),-1/alpha) #to dist, FF, to contact in one line

Thank you.