search

zhxiaokang / RASflow

RNA-Seq analysis workflow
MIT License
103 stars 57 forks source link

multiQC conflicts in conda install #39

Open EamonCoughlan opened 7 months ago

EamonCoughlan commented 7 months ago

MultiQC doesn't appear to run in the current state of the conda environment from the yaml file. Attempting to run on a minimal dataset (2 samples) throws 'SyntaxError: future feature annotations is not defined', which seems like it may be related to use of Python version <3.7 . This is overcome by using Python 3.7, but then copytree requires 3.8, which is incompatible with other dependencies.

This is occurring in Linux on both a HPC and a local WSL install.

zhxiaokang commented 7 months ago

Hi, that's very strange, since if your environment is created with the yaml file, every tool should work properly with each other. As you tried, if Python is upgraded, it may lead to other conflicts. Could you try to use the Docker image? Inside the docker image, everything is fixed and the version conflict should not be there.

EamonCoughlan commented 7 months ago

The docker image works. I was hoping to avoid it since our HPCC doesn't support Docker but maybe I can just run everything on my machine instead.

zhxiaokang commented 7 months ago

I just tried to create the conda env with the yaml file but failed. It seems that some old-version packages are not maintained properly. The workflow definitely needs a thorough update, but it's unfortunately not on my priority list. So currently the compromised solution would be using the docker image.

bpickett commented 3 weeks ago

I am running a transcriptome analysis using the full-length versions of the 4 example fastq files. I get the following error from the rule DEA:

Error in $<-.data.frame(*tmp*, "subject", value = integer(0)) : replacement has 0 rows, data has 4 Calls: DEA -> $<- -> $<-.data.frame

Execution halted

Any ideas about what might be happening and how to fix it?