smiles2pdb

smiles2pdb is a small command-line tool to convert SMILES string into PDB files and optionally generate different conformations of the small molecules.

Why C++?

This tool uses RDKit libraries and according to their own documentation most functionalities of the C++ libraries can be used via the Python API.

The main purpose of this tool here is to be a component of the WeNMR service portal to handle the generation of PDB files and for that we need to have a small and portable executable.

By using the C++ libraries directly, we can create a statically compiled binary with no system dependencies that can be executed anywhere.

Usage

$ smiles2pdb
Usage: smiles2pdb <SMILES_string> <output_file.pdb> [num_conformers]
  num_conformers: Number of conformers to generate (default: 1)

For example:

smiles2pdb "Cc1c(nc2ccc(F)cc2c1C(O)=O)c3ccc(cc3)c4ccccc4" 1D3G_BRE.pdb 10

The resulting 1D3G_BRE.pdb will be a multi-model ensemble containing 10 different conformations.

TODO

• Tune RMSD cutoff between conformations
• Add support for SELFIES

Compilation

Compiling this tool as a static binary is non-trivial since it requires several system packages and a from source installation of RDKit.

To facilitate the compilation we have provided a Dockerfile that acts a builder.

Compile the tool inside the container with (this operation will take several minutes):

docker build -t builder .

Copy the compiled binary to the current working directory with

docker run \
  --rm \
  -u $(id -u):$(id -g) \
  -v $(pwd):/output \
  builder \
  cp /app/build/smiles2pdb /output/

Execute it:

./smiles2pdb