bhattacharya-lab/piqle

Protein-protein interface quality estimation by deep graph learning

PIQLE: protein-protein interface quality estimation by deep graph learning of multimeric interaction geometries

by Md Hossain Shuvo, Mohimenul Karim, Rahmatullah Roche, Debswapna Bhattacharya

published in Bioinformatics Advances

Codebase for our protein-protein interface quality estimation by deep graph learning, PIQLE.

Environment

PIQLE is tested on x86_64 Linux system in the following Python environment

1. python 3.6.13
   
2. dgl 0.9.0
   
3. torch 1.10.2
   

Download and installation

$ git clone https://github.com/Bhattacharya-Lab/PIQLE.git
$ cd PIQLE
$ python config.py

Usage

To see usage instructions, run

$ python PIQLE.py

You should see the following output

usage: PIQLE.py [-h] [--tgt TARGETNAME] [--seq FASTAFILE] [--dec DECOYDIR]
                [--ch CHAINFILE] [--msa1 INMSA1] [--msa2 INMSA2]
                [--a3m1 INA3M1] [--a3m2 INA3M2] [--out OUTDIR]

Arguments:
  -h, --help        show this help message and exit
  --tgt TARGETNAME  Target name
  --seq FASTAFILE   Fasta file
  --dec DECOYDIR    Complex decoy directory
  --ch CHAINFILE    Chain file
  --msa1 INMSA1     MSA1: Multiple Sequence Alignment of chain 1
  --msa2 INMSA2     MSA2: Multiple Sequence Alignment of chain 2
  --a3m1 INA3M1     MSA1 in A3M format
  --a3m2 INA3M2     MSA2 in A3M format
  --out OUTDIR      Output directory.

Example commands to run PIQLE

python PIQLE.py --tgt example --seq absolute_path/example/example.fasta --dec absolute_path/example/decoys/ --ch absolute_path/example/example.chain --msa1 absolute_path/example/msa/example_A.aln --msa2 absolute_path/example/msa/example_B.aln --a3m1 absolute_path/example/msa/example_A.a3m --a3m2 absolute_path/example/msa/example_B.a3m --out absolute_path/example_out/

Please see the output example/example.PIQLE of the above example command.

Argument descriptions

IMPORTANT: Please provide the absolute paths for all the input files including the output directory

• --tgt TARGETNAME: name of the target without any extension
  
• --seq FASTAFILE: fasta file containing both of the sequences of the complex. Should be exactly in the same order as the chain order in the PDB. Please see example/example.fasta
  
• --dec DECOYDIR: directory containing ONLY complex pdb files. Please see example/decoys
  
• --ch CHAINFILE: chain identifier of the complex. Only one line with chain ids seperated by space. Should be exactly in the same order as the chain order in the PDB. Please see example/example.chain
  
• --msa1 INMSA1: multiple sequence alignment (e.g. '.aln file') of the first chain in the complex. Please see example/msa/example_A.aln
  
• --msa2 INMSA2: multiple sequence alignment (e.g. '.aln file') of the second chain in the complex. Please see example/msa/example_B.aln
  
• --a3m1 INA3M1: A3M of chain1 (e.g. '.a3m file') of the first chain in the complex. Please see example/msa/example_A.a3m
  
• --a3m2 INA3M2: A3M of chain2 (e.g. '.a3m file') of the second chain in the complex. Please see example/msa/example_B.a3m
  
• --out OUTDIR: output directory. IMPORTANT: an output directory will be created, if it does not already exist.

piqle's Issues

files not found!

I have problem while I want to run for examples: this is my error, despite I am sure that I have datasets, but I have directory error.

cp: cannot stat 'example/msa/example_B.a3m': No such file or directory
python /home/fe5vb/project/PIQUE/PIQLE//example/msa/msa_concat.py --tar example --rec example/msa/example_A.a3m --lig example/msa/example_B.a3m --ch1 A --ch2 B --out ./
/home/fe5vb/project/PIQUE/PIQLE//apps/glinter/external/A3M_NoGap example/msa/example_A.a3m .//example/msa/example_A.a3m.nogap
file example/msa/example_A.a3m not found!
file example/msa/example_B.a3m not found!

specbloc not found

Hello!

I was trying to perform a docking analysis between two proteins and I keep running into the error that the corresponding specbloc file of the first a3m file isn't found. How can I remedy this?

Best,
Jasper

a3m and aln files

thank you for your good repo.
I have only pdb and fast files, how can I have .aln and .a3m files?
is there any specific software to generate the exact format as your data format?