InChI and SMILES stand as the two de facto standard line notations for cheminformatics. A line notation represents a molecular graph as a compact string of characters without newlines. Given its similarity to SMILES, a reasonable assumption would be that InChI can be parsed to regenerate the molecular graph that created it. With few exceptions, this is not the case. This article explains why.

Simplified Molecular Input Line Entry System (SMILES) is the de facto standard for compact molecular representation in cheminformatics. Due to this role, users have come to expect SMILES support in the software packages they use. Developers have responded to the demand with a wide range of products capable of reading and writing SMILES. These products form a broad network of tools that can in principle work together through a common language.

SMILES is the de facto standard for compact molecular representation in cheminformatics. One reason for its efficiency is hydrogen suppression, which has been covered here before. Of course, the thing about leaving hydrogens out of a molecular graph is that sooner or later they’ll need to be put back in. There are many ways to do this, with varying costs. This article takes a look at this problem and proposes a simple, efficient solution.

Iterators are ubiquitous in Rust and mostly just work. Even so, Rust Iterators carry some noteworthy caveats. For an example, consider the case of returning an Iterator from a method defined on a trait. This article describes the problem and offers evidence supporting a practical and ergonomic solution. A Graph Trait Imagine designing a Rust crate for working with graphs. Here I’m referring to the mathematical construct, not a chart.

Early-phase medicinal chemistry is full of challenges and uncertainties. But despite it all, the entire enterprise can often be boiled down to just one question: which compounds should be tested next? Opinions differ on the role that cheminformatics should play here. One school of thought says to simply supply the answer. Recent advances in machine learning have bolstered this view.

A previous article presented a formal grammar for the SMILES language. Its most notable feature was the lack of left-recursion, which enables simple, efficient parsers to be written. However, that first attempt left some room for improvement. In this article, a new grammar is presented that addresses outstanding points. Limitations of Previous Grammar Limitations of the previous grammar included: Multiple unbounded quantifiers.

SMILES is the de facto standard for information exchange in cheminformatics. But despite this language’s prominent role, low-level tools and documentation remain sparse and incomplete. Recent articles here have highlighted the development of a SMILES grammar and a parser implementation in Rust. This article extends that theme by introducing another powerful tool for the analysis and manipulation of SMILES using Rust.

Machine learning has made major inroads into molecular property prediction over the last few years. But with the deluge of techniques and tools comes an increased need to recast what works well into a form usable by non-experts. DeepChem is a toolchain designed to do just that. This article presents a short, simple code example demonstrating how to use DeepChem to train and use a deep convolutional network to predict aqueous solubility.

The Rust homepage cites “blazingly fast” performance as an important reason to consider the language, and top-tier performance is often cited by developers as a reason they chose Rust. So far, this blog has emphasized Rust’s many other compelling features. This article breaks from that focus to examine performance. Specifically, it compares the speed with which RDKit (Python/C++) and ChemCore (Rust only) can read SMILES.

Electrons don’t get much respect in cheminformatics. Although atoms take star billing in everything from file formats to toolkits to applications, electrons sometimes don’t even make the credits. When electrons do find themselves on the cast, they tend to play offstage characters; they are the Mrs. Wolowitz of cheminformatics, sometimes heard but never seen.

SMILES is an industry standard with a checkered standardization history. Both primary literature descriptions and online documentation left crucial points unaddressed. These limitations resulted in OpenSMILES, the first independent attempt to codify the syntax and semantics of the SMILES language.