RDMC Documentation#

_images/RDMC_icon.svg

Introduction#

RDMC (Reaction Data and Molecular Conformer) is an open-source lightweight software package specialized in handling Reaction Data and Molecular (including transition states) Conformers.

It contains various modules and classes helpful for relevant tasks to make conversion, visualization, manipulation, and analysis of molecules easier. rdmc and rdtools are the two major modules.

Installation#

The source code of the RDMC software package is hosted on GitHub, and its binary distribution is available on Anaconda Cloud and PyPI. The easiest way to install RDMC is to use conda or mamba:

conda install -c xiaoruidong rdmc  # replace conda to mamba to use mamba

Or

pip install rdmc

conda can be installed by via Anaconda, and mamba can be installed via Mambaforge.

You can also install RDMC from the source code:

git clone https://github.com/xiaoruidong/rdmc
cd RDMC
conda env create -f environment.yml
conda activate rdmc
python -m pip install --no-deps -vv ./

rdtools#

It has a collections of functions that can be directly operated on to RDKit native objects. Most of them can be regarded as a modified version of a RDKit native operation but with simplified imports, more intuitive usage, and more robustness implementation. They are distributed in submodules, each named by its relevance field and objects. It is best used in cases where efficiency is a major conern and the task you required is relatively simple. Here are some highlights in rdtools:

  • viewers in rdtools.view. These viewers greatly extend RDKit’s default Ipython 3D viewer, with the capability of viewing animations and interaction with conformers.

  • generate_resonance_structures in rdtools.resonance is able to generate resonance structures for radical molecules that is not capable by the original RDKit.

  • mol_from_smiles in rdtools.conversion make sure the created molecule has an atom ordering consistent with the atom mapping

  • mol_from_xyz supports two backends openbabel as well as xyz2mol for molecule connectivity perception. If both native backends fail (e.g., cannot be sanitized, or wrong charge or multiplicity), rdtools also provided a heuristic fix tool fix_mol in rdtools.fix to help fix the molecules if possible.

rdmc#

It can be regarded as a midware between RDKit/rdtools basic operations and complicated workflows. Mol (Previously, RDKitMol) and Reaction are the most important classes.

  • Mol (known as RDKitMol previously) is a child class of RWMol, which means that you can directly use it with RDKit’s native functions, but also integrated a lot of tools in rdtools, so you can directly use them as class methods. The appended methods not only provides convenience in usage, but also make sure the output molecule objects, if applicable, is still a rdmc.Mol object. While many RDKit functions will just output Chem.Mol which is no longer editable, breaking the flow of your molecule operations.

  • Reaction provides intuitive APIs for getting bond analysis, reaction comparison, visualization, etc.

We provide examples of how to combine rdtools and rdmc with other dependencies to build useful tools. Python native interactive Log file parsers in rdmc.external.logparse are a good show case where rdmc and cclib are combined. We also provide solutions to pipelining tasks to achieve high-throughput generating and processing of large amount of molecule/reaction data in rdmc.conformer_generation.

Requirements#

RDMC is written in Python (>= 3.7) and has dependencies only on popular packages.

  • To use rdtools, you only needs numpy and rdkit at minimum. You can install optional dependencies: scipy for better resonance structure generation for polycyclic molecules, py3dmol to use the amazing 3D viewers, openbabel to extend rdmc’s xyz perception cability.

  • To use rdmc, the dependencies are basically the same as rdtools, but we do recommend installing all optional dependencies for a better experience. Besides, to plot curves and figures for data, you can install matplotlib; to play around with the log parsers you should consider install cclib and ipywidgets. And to start computations in conformer_generation, you need to have xtb and orca (which are free to academia) installed to get some serious results.

But in a word, RDMC’s dependencies are very general-purpose and popular packages.

First Run#

To start with, simply try the following:

from rdmc import Mol, Reaction
mol = Mol.FromSmiles('CCO')
rxn = Reaction.from_reaction_smiles('CCO>>CC(=O)O')

And see what the mol and rxn are capable of! The full lists of APIs of Mol and Reaction are provided in this documentation.

Examples#

We also provided a few notebooks (available in \ipython and Colab) to demonstrate the usage of RDMC. Please feel invited to try them out!

Developers#

  • Dr. Xiaorui Dong ( | | ),

  • Dr. Lagnajit Pattanaik ( | | ),

  • Dr. Shih-Cheng Li ( | | ),

  • Dr. Kevin Spiekermann ( | | ),

  • Dr. Hao-Wei Pang ( | | ),

  • Jonathan W. Zheng ( | | ),

  • Prof. William H. Green ( | )

at Green Research Group at Massachusetts Institute of Technology (MIT). For any questions while using RDMC, please contact us via the GitHub issue page or email us at rdmc_dev@mit.edu.

Contents#

APIs#