search

lanagarmire / deepimpute

An accurate and efficient deep learning method for single-cell RNA-seq data imputation
MIT License
84 stars 27 forks source link

deepImpute multinet division by 0 issue #1

Closed jk86754 closed 5 years ago

jk86754 commented 6 years ago

Hi,

Trying to run deepimpute under python 3.6.4 i ran into the following issue:

Traceback (most recent call last): File "/sc/wo/home/karmajx/jobs/deepimpute_slephase1.py", line 4, in imputed=deepImpute.deepImpute(data, NN_lim='auto', cell_subset=1) File "/sc/wo/app/python/python3/deepimpute/deepimpute/deepImpute.py", line 17, in deepImpute multi.fit(data, NN_lim=NN_lim, cell_subset=cell_subset) File "/sc/wo/app/python/python3/deepimpute/deepimpute/multinet.py", line 130, in fit n_runs, n_cores = self._getRunsAndCores(numberOfTargetGenes) File "/sc/wo/app/python/python3/deepimpute/deepimpute/multinet.py", line 103, in _getRunsAndCores self.NN_params["n_cores"] = max(1, int(self.maxcores / n_cores)) ZeroDivisionError: division by zero

help would be appreciated.

Thanks,

Jozsef

breckuh commented 6 years ago

Hi Jozsef,

Thanks for reporting this and sorry about the trouble. I have just pushed an update that should give us more diagnostic info on the shape of your input data. If you get a moment to pull the latest and rerun, the section that would help us debug looks like this:

.Input dataset is 3000 genes and 500 cells
First 3 rows and columns:
                  ENSG00000177954  ENSG00000197756  ENSG00000231500
AATTGTGACTACGA-1            826.0            674.0            694.0
TGACACGATTCGTT-1            617.0            618.0            594.0
TGTCAGGATTGTCT-1            525.0            550.0            540.0

Thanks!