Running Pangolin

Question

[Sruthi]

Hello,

I would like to use Pangolin splicing prediction for my data. However, I face issues on running the commands from their GitHub.

So, this is the Pangolin GitHub site: https://github.com/tkzeng/Pangolin/blob/main/README.md

I have installed torch, pangolin, other dependencies such as:

conda install -c conda-forge pyvcf
pip install gffutils biopython pandas pyfastx

and have my input files ready in this format:

chr1    100017682   G   A
chr1    100059878   G   T
chr1    100140182   G   C
chr1    100273394   G   T

So, now I run this command:

pangolin input.txt gencode.v43.pc_transcripts.fa gencode.v38.annotation.db output

It gives me an error saying gffutils module not found.

Traceback (most recent call last):   File
"/Users/sruthisrinivasan/miniconda3/bin/pangolin", line 5, in <module>
    from pangolin.pangolin import main   File "/Users/sruthisrinivasan/miniconda3/lib/python3.10/site-packages/pangolin/pangolin.py",
line 5, in <module>
    import gffutils ModuleNotFoundError: No module named 'gffutils'

But I have gffutils installed.

pip3 show gffutils
Name: gffutils
Version: 0.12
Summary: Work with GFF and GTF files in a flexible database framework
Home-page: https://github.com/daler/gffutils
Author: Ryan Dale
Author-email: [e-mail]
License: 
Location: /Users/sruthisrinivasan/miniconda3/lib/python3.10/site-packages
Requires: argcomplete, argh, pyfaidx, simplejson, six

Could you help me understand what is the issue here?

Thank you!

[GenoMax]

Are you following the instructions from the site you linked? First step is to create an annotation database.

If that is done the next step says

Run Pangolin on a VCF or CSV file containing a list of variants.

Your input.txt is not a VCF/CSV file.