Try to generate a cyclic peptide from an amino acid sequence with Python and RDKit

Introduction

In order to realize the subject, I found a library called CycloPs, so I tried it this time.

environment

• Python3.x
• RDKit 2020.X
• CycloPs

What are CycloPs?

It is a tool that generates a virtual library of cyclic peptides from amino acid sequences. By binding residues to each other from a one-dimensional amino acid sequence, a cyclic peptide is produced. It can be used both on the GUI and on the command line. See the bibliographic treatise for details.

There are five types of patterns that bind residues to each other. It seems that each has detailed conditions, so please see the paper for details.

• head-to-tail bond It binds the C- and N-terminus of the amino acid sequence.
• disulfide bond It binds cysteine ​​residues to each other.
• side-chain to side-chain bond Join the side chains together.
• side-chain to N-terminal bond Join the side chain and the N-terminus
• side-chain to C-terminal bond Join the side chain and C-terminus

Install CycloPs

Originally it was developed in Python 2 series, but since RDkit did not work well in my environment, I installed it in 3 series.

　$ conda create -n cyclops python==3.6
　$ conda activate cyclops
　$ conda install -c rdkit rdkit
  $ git clone https://github.com/fergaljd/cyclops.git

Post-installation work

Since CycloPs does not work in Python3 series, it is necessary to modify the source. This time, you only need to use the command line. Therefore, modify `PepLibGen/StructGen/StructGen.py` as follows to make it work with Python3.

Fixes

--Modify the description such as print a, b to print (a, b). --Correct the description such as Except a, b to Except a as b. --An error occurs around line 940, so fix it as follows.

            # print(out,mollstr)
            # print(out, '$$$$')
            out.write(molstr)
            out.write('$$$$')
            out.write('\n')

--Fix other runtime errors to work with Python 3 (cut out).

Execution example

Execute by specifying the amino acid sequence and output SDF of the cyclic peptide registered in CPPsite 2.0. Then, the list of cyclic peptides produced by the five patterns is output to SDF. The following is an execution example.

python PepLibGen/StructGen/StructGen.py CSKSSDYQC hoge.sdf

Execution result example

The six structures obtained by specifying the amino acid sequence "CSKSSDYQC" are listed below.

First

This is a linear type. It seems that the specifications are such that non-annular ones are also output.

The second

Is this a connection between the N-terminus and the C-terminus?

Third

Is this the end and side chain?

4th

Is this a side chain?

Fifth

Is this a side chain?

6th

This would be a bond between cysteine ​​residues.

reference

• CycloPs: generating virtual libraries of cyclized and constrained peptides including nonnatural amino acids
• CycloPs: A Cyclic Peptide Library Generator
• CPPsite 2.0