jinghuanglab / slcfold Goto Github PK

The source code is the implementation of the fitness score, which was designed for new folds. Through ranking by the score, the new fold SLC44 was successfully discovered from the human SLC superfamily. The fitness score can also be adopted to other families for new folds.

1. 3D structures from multiple prediction methods. RaprorX-3DModeling and trRosetta were used in the paper.
2. Contact maps. TripleRes was adopted.
3. Secondary structures. DeepCNF was adopted.

Files in input directory:

├── SLC44A1                           # the name of target protein
│   ├── SLC44A1.fold.229863005.pdb    # predicted structure by RaptorX-3DModeling
│   ├── SLC44A1.partial.contact       # predicted contact by TripletRes
│   ├── SLC44A1.partial.fasta         # sequence
│   ├── SLC44A1.partial.pdb           # predicted structure by trRosetta
│   └── SLC44A1.ss3                   # secondary structure by DeepCNF
└── SLC6A12
    ├── SLC6A12.fold.361676800.pdb
    ├── SLC6A12.partial.contact
    ├── SLC6A12.partial.fasta
    ├── SLC6A12.partial.pdb
    └── SLC6A12.ss3

Note: The suffixes of the three kinds of predicted information need to be set as "pdb", "contact", and "ss3".

Structure from RCSB PDB in PDB format. Here, only a few pdb files are included in PBDs. All membrane protein structures considered in the paper are listed in memPDBs.

The tool TMalign and TMscore are required. And the corresponding paths of executables should be set in scripts/ranking_score.py.

The variable NUMPROCESS in scripts/ranking_score.py can be set according to the number of cpu availiable in your machine.

The code is implemented in python3.6. The required packages are as follows:

glob
pandas
numpy
biopython

cd scripts/
python ranking_score.py --input ../input/ --term   # get each term of the score for all proteins
python ranking_score.py --input ../input/ --score  # get the score values for all proteins

The results are in output.

Tengyu Xie, Ximin Chi, Bangdong Huang, Fangfei Ye, Qiang Zhou, Jing Huang. Rational Exploration of Fold Atlas for Human Solute Carrier Proteins. bioRxiv, 2021.08.21.457230; doi: https://doi.org/10.1101/2021.08.21.457230.