Jeremy Monat

As the YouTubers would say, “A lot of you have been asking me about how to write cheminformatics blog posts.” Well, not a lot, but at least a couple! I realized that there’s a pattern to how I write cheminformatics blog posts (16 so far), so I’m sharing that here. My blog posts are intended to be tutorials that explore a topic, usually with existing tools. I figure out how to accomplish a cheminformatics objective, then share that.

How to Write Cheminformatics Blog Posts

Tautomers are chemical structures that readily interconvert under given conditions. For example, an amino acid has a neutral form, and a zwitterionic form with separated positive and negative charges. Cheminformatics packages have algorithms to enumerate tautomers based on rules. Which algorithms produce the most tautomers? And how successful is InChI at representing with a single representation all tautomers of a given structure?

Comparing Tautomer Generation Algorithms

I contributed MolsMatrixToGridImage to the RDKit 2023.09.1 release because I found myself writing similar code over and over to draw row-and-column grids of molecules. For projects where each row represented something, such as a molecule and the fragments off a common core, my mental model corresponded to a two-dimensional (nested) data structure, whereas the pre-existing function MolsToGridImage supported only linear (flat) data structures.

MolsMatrixToGridImage Simplifies Code

This blog post presents a more computationally-efficient way to determine the abundance of the molecular isotopes of a molecule. In part 1, we created a molecule for each possible placement of each isotope in a molecule. While that worked, it was computationally expensive because it required creating each permutation. In this blog post, we’ll create each combination only once and calculate its abundance using the binomial distribution.

Molecular Isotopic Distributions Take 2: Combinations

Harry Wiener was “a pioneer in cheminformatics and chemical graph theory”. In his 1947

 Journal of the American Chemical Society

article “Structural Determination of Paraffin Boiling Points”, he introduced the path number $\omega$ “as the sum of the distances between any two carbon atoms in the molecule, in terms of carbon-carbon bonds”, which is now known as the Wiener index.

Revisiting a Classic Cheminformatics Paper: The Wiener Index

Exploring other cheminformatics toolkits besides the RDKit, I wanted to try EPAM Indigo Toolkit. The Indigo Toolkit is free and open-source with Apache License 2.0, so it can be used in proprietary software. I was unable to find simple examples of drawing molecules in a Python Jupyter Notebook, so here’s how to do that. This post also demonstrates how to save molecular images to a file.

Drawing molecules with EPAM’s Indigo cheminformatics library

Here’s how to display formatted molecular formulas in tables and graphs. In addition to formatted molecular formulas, these techniques should work for any Markdown or LaTeX. In the last blog post, we generated molecular formulas from SMILES strings or RDKit molecules. Once we have those molecular formulas, formatted as Markdown or LaTeX, we might want to display them in tables or graphs.

Displaying Molecular Formulas in Molecular Grids, Tables, and Graphs for Elemental Analysis

In cheminformatics, the typical way of representing a molecule is with a SMILES string such as CCO for ethanol. A SMILES string can be converted into a molecular graph, which can be used to determine molecular structure and related properties. However, there are still cases where the molecular formula such as C

 2

H

 6

O is useful.

Molecular Formula Generation

In a previous post, I revisited Wiener’s paper predicting alkanes’ boiling points using modern cheminformatics tools. This follow-up post refits the data with modern mathematical tools to check how well the literature parameters, and the current parameters optimized here, fit the data. Wiener and Egloff’s works are impressive for using cheminformatics parameters that model physical data with simple relationships.

Refitting Data From Wiener’s Classic Cheminformatics Paper

This utility plots a mass spectrometry fragmentation tree given the species’ SMILES and their hierarchy, that is which species fragments into which other species. Here’s an annotated example.

Draw a Mass Spectrometry Fragmentation Tree Using RDKit

In drug discovery, the lead optimization step often involves creating analogues of a hit (a promising compound which produces a desired result in an assay) to optimize selectivity and minimize toxicity. Because it is typically easier to chemically modify the periphery of the molecule (for example the functional groups) than the scaffold, it is helpful to compare the groups off of the common scaffold.

Find the Maximum Common Substructure, and Groups Off It, For a Set of Molecules Using RDKit

This example uses machine learning to predict the lipophilicity of compounds. Lipophilicity measures how well a compound dissolves in non-polar media such as fats and lipids. So it’s important for drugs that are delivered orally (for example, via a pill) because the active ingredient needs to be absorbed into the lipids of biological membranes.

Chemistry machine learning for drug discovery with DeepChem

This notebook is the result of a collaboration between myself, Dr. Jeremy E. Monat, and Dr. Phyo Phyo Zin. This is a hands-on tutorial that guides readers through implementing skeletal editing transformations using Python and RDKit, and Phyo Phyo focuses on the high-level principles and applications in the blog post How to Use Skeletal Editing in Drug Discovery: Key Benefits &amp; Applications.

Tutorial for Skeletal Editing in Drug Discovery

Elements can have several isotopes, which have the same number of protons and electrons but different numbers of neutrons. Because a neutron has a mass of approximately 1 amu (atomic mass unit), different isotopes of an element appear at different mass-to-charge ratios in a mass spectrum as measured by a mass spectrometer.

Molecular Isotopic Distributions Take 1: Permutations

This example lets you visualize the hierarchy of nonbinary trees. An example is the classification of chemical isomers, which are compounds that have the same chemical formula, but different arrangements of atoms in space.

Visualizing Nonbinary Trees: Classification of Chemical Isomers

This utility reports whether the starting materials are commercially available for a set of synthesis targets given reactions. You give it your synthesis targets and the reaction to create each, it determines the starting materials, checks whether they are commercially available, and tells you whether each target is accessible–whether all its starting materials are commercially available.

Are the Starting Materials for Synthesizing Your Target Molecules Commercially Available?

Retrosynthetic analysis involves decomposing a target molecule into a set of fragments that could be combined to make the parent molecule using common reactions. The Recap algorithm by X. Lewell, D. Judd, S. Watson, and M. Hann accomplishes that. Recap is implemented in the RDKit cheminformatics Python package.

Retrosynthetically Decompose a Molecule In a Tree Structure Using Recap in RDKit

Molecules have a color if their electronic energy levels are close enough to absorb visible rather than ultraviolet light. For organic molecules, that’s often because of an extensive chain of conjugated bonds. Can we use cheminformatics to find evidence that increasing conjugated bond chain length decreases absorption wavelength, which makes a molecule colored?

Why some organic molecules have a color: Correlating optical absorption wavelength with conjugated bond chain length

When analyzing a set of molecules, you might want to find the maximum common substructure (MCS) match between them. This utility function SmilesMCStoGridImage does that for a set of molecules specified by SMILES, displays the SMARTS substructure as a molecule, and displays all the molecules in a grid with that substructure highlighted and aligned.

Find and Highlight the Maximum Common Substructure Between a Set of Molecules Using RDKit

An LLM Chat Plants the Seed   To come up with small projects to do in preparation for job interviews (contact me if you need a cheminformatics scientific software developer!), I interactively asked ChatGPT for suggestions. It suggested The Murcko Scaffolds Blog Post   I wrote code to find the Murcko scaffold for a set of molecules, then determine the count of unique scaffolds. That answered the ChapGPT challenge.

Murcko Scaffolds Post on the RDKit Blog: An LLM Circle

This tutorial shows two ways to accomplish skeletal editing reactions in the RDKit: By using Reaction SMARTS or atom editing. As my collaborator Dr. Phyo Phyo Zin posted in How to Use Skeletal Editing in Drug Discovery: Key Benefits &amp;

Skeletal Editing with Reaction SMARTS and RDKit Atom Substitution

This is an update of a Recap retrosynthetic tree blog post to programmatically annotate the diagram. Previously, I exported the drawing from the RDKit and manually annotated the it with another program. After reading Greg Landrum’s RKDit blog post Drawing on drawings, I wanted to automate annotating the tree.

Jeremy Monat, PhD

Retrosynthetically Decompose a Molecule In a Tree Structure Using Recap in RDKit

Find and Highlight the Maximum Common Substructure Between a Set of Molecules Using RDKit