/DEPECT

Design and Engineering of Protein and Enzymes by Computational Tools

Primary LanguageJupyter NotebookMIT LicenseMIT

DEPECT

Design and Engineering of Proteins and Enzymes by Computational Tools

Introduction

Enzymes are key to bioindustry and some proteins are very important like COVID-19 neutralize antibody. However, it is hard to get efficient or useful biomacromolecules. Here I present DEPECT package for more easily and faster in silico design of biomacromolecules.

Click the link to download the linux executable binary file of hybrid_redesign. It has been tested under Ubuntu 20.04.

To run hybrid_redesign scripts, make sure you have Rosetta, muscle and ncbi-blast+ suite installed!

usage:

./hybrid_redesign -s 4eb0.pdb -cid A -db /ndata/databases/blastdb/nr -ic 30 -e 1e-20 -sc 2 -nt 16 -m blast

After the calculation done(You can monitor it via top), run hrd.sh to generate combined hybrid_design.sc file.

combine autodock suites and modeller to design enzymes.

To run enzde scripts, make sure you have AutoDock suites and modeller installed!

  • Install
    Edit the depect_enzde_config.py, change the ADT_PATH to where its installed.
  • Demo
    ./demo.sh

Doc

mutfile format

12|LYS|A  
|  |   |  
|  |   |___Chain ID  
|  |___Mutate to which AA (Here is to mutate to a LYS)   
|___Residue Sequence Number in Pdb File   

cstfile format

CST_NAME:atom1,atom2
If the atom in the protein:
protein|chain id|number of residue|atom name
If the atom in the ligand:
ligand|atom number

in the zlj.cst used in demo:

distace1:protein|A|79|OG,ligand|13  

If the constraint is an angle, use 3 atoms, for a dihedral use 4 atoms.