Exploratory Data Analysis With mols2grid and Bemis-Murcko Frameworks

One of the most common tasks in Cheminformatics is exploratory data analysis (EDA).  Given a new dataset, we often need to rapidly explore the chemistry in a set containing hundreds, or even thousands, of molecules.  One useful technique for EDA is the Bemis-Murcko framework.  This technique, originally published by Guy Bemis and Mark Murcko, provides a simple but elegant means of grouping molecules.  Bemis-Murcko frameworks (also known as scaffolds) are created by successively removing monovalent atoms until only ring atoms and linker atoms remain.  There are a few nuances having to do with the removal of exocyclic double bonds and the maintenance of aromaticity, but the method itself is very easy to understand.  There are two versions of the Bemis-Murcko framework, which are sometimes confused.  In the first version, illustrated in the top row of the figure below, monovalent atoms are removed until only ring atoms and linker atoms remain.  In the second version, a generic framework is generated by transforming the result of the previous procedure by converting all atoms to carbon atoms and all bonds to single bonds.  

While both methods can be useful,  I tend to use the first method, which maintains the atom types and bond orders from the input structure.   

In the Jupyter notebook associated with this post, we take a set of structures from the ChEMBL database, calculate the Bemis-Murcko frameworks, then use mols2grid to display the frameworks.  By itself, this would be kind of cool, but Cedric Bouysset made some updates to mols2grid to make this even cooler.  In the latest version of mols2grid, we can create a callback function that is called whenever the user clicks on a structure in the grid.  In his talk at the RDKit User Group Meeting, Cedric showed a nice example where clicking on a molecule in the grid displays a 3D structure below the grid using py3Dmol.  I modified Cedric's code a bit to enable a click on the grid to display another grid.  This provides an easy way to look at things like scaffold->molecule or cluster center->cluster member relationships.  



The notebook also shows how we can use SMARTS to search the grid or the underlying dataframe.   I've found this sort of workflow to be incredibly useful when I'm exploring a new dataset or trying to understand a set of assay results.  Hopefully, you'll find this useful too. 

The Jupyter notebook can be found here on GitHub or can be run directly on Google Colab. 





Comments

Post a Comment

Popular posts from this blog

We Need Better Benchmarks for Machine Learning in Drug Discovery

Image
Most papers describing new methods for machine learning (ML) in drug discovery report some sort of benchmark comparing their algorithm and/or molecular representation with the current state of the art.  In the past, I’ve written extensively about statistics and how methods should be compared .  In this post, I’d like to focus instead on the datasets we use to benchmark and compare methods.  Many papers I’ve read recently use the MoleculeNet dataset, released by the Pande group at Stanford in 2017, as the “standard” benchmark.   This is a mistake.  In this post, I’d like to use the MoleculeNet dataset to point out flaws in several widely used benchmarks.  Beyond this, I’d like to propose some alternate strategies that could be used to improve benchmarking efforts and help the field to move forward.   To begin, let’s examine the MoleculeNet benchmark, which to date, has been cited more than 1,800 times.   The MoleculeNet collection consists of 16 datasets divided into 4 categories.  Qua
Read more

AI in Drug Discovery 2023 - A Highly Opinionated Literature Review (Part I)

Image
Here’s the first part of my review of some interesting machine learning (ML) papers I read in 2023.  As with the previous editions , this shouldn’t be considered a comprehensive review.  The papers covered here reflect my research interests and biases, and I’ve certainly overlooked areas that others consider vital.  This post is pretty long, so I've split it into three parts, with parts II and III to be posted in the next couple of weeks.    I. Docking, protein structure prediction, and benchmarking II. Large Language Models, active learning, federated learning, generative models, and explainable AI III. Review articles 2023 was a bit of a mixed bag for AI in drug discovery.  Several groups reported that the deep learning methods for protein-ligand docking weren’t quite what they were initially cracked up to be.  AlphaFold2 became pervasive, and people started to investigate, with mixed success, the utility of predicted protein structures.  There were reports of significant advanc
Read more

Getting Real with Molecular Property Prediction

Image
Introduction If you believe everything you read in the popular press, this AI business is easy. Just ask ChatGPT, and the perfect solution magically appears. Unfortunately, that's not the reality. In this post, I'll walk through a predictive modeling example and demonstrate that there are still a lot of subtleties to consider. In addition, I want to show that data is critical to building good machine learning (ML) models. If you don't have the appropriate data, a simple empirical approach may be better than an ML model.  A recent paper from Cheng Fang and coworkers at Biogen presents prospective evaluations of machine learning models on several ADME endpoints.  As part of this evaluation, the Biogen team released a large dataset of measured in vitro assay values for several thousand commercially available compounds.  One component of this dataset is 2,173 solubility values measured at pH 6.8 using chemiluminescent nitrogen detection (CLND), a technique currently consider
Read more