some issues found so far

[zhaoshuoxp]

Hi, this is Quanyi from the characterization group (Jesse), Quertermous lab. I am working on running the pipeline with the TF_Perturb_Seq_Pilot dataset, but I have encountered several issues:

1. Error with downloadGenome Process
   Process 'prepare_mapping_pipeline:downloadGenome' declares 2 input channels but 1 was specified.
   It seems the downloadGenome function in ./processes/downloadGenome.nf requires two inputs. I commented out the line val hg38_path to proceed.

2. Parameter Issue in prepare_user_guide_inference()
   In process_mudata_pipeline.nf and process_mudata_pipeline_HASHING.nf, the function prepare_user_guide_inference() requires the path to the mudata file generated in the previous step. The correct parameter is Guide_Assignment.guide_assignment_mudata_output, not file(params.input) on lines 60, 66, and 72.

3. GTFProcessing Installation Issue
   The GTFProcessing package was not correctly installed in the Conda environment, despite being listed in conda_envs/nextflow.yaml. To fix this, I manually installed the tool in the Conda environment located at ./work/conda/spec-file-xxxxxxx, which enabled the required Python scripts in the bin/ directory to work properly. --- Not sure if this issue is specific to me.

4. Issue in bin/prepare_inference.py
   On line 25 of bin/prepare_inference.py, the script extracts gene symbols for intersections. However, the example dataset uses Ensembl IDs in the user_pairs_to_test.csv file. I changed it to use gene symbols, which resolved the issue.

I hope these will be helpful to the others.