A recent publication caught my eye.
Extraction of organic chemistry grammar from unsupervised learning of chemical reactions DOI.
Anyone who has been involved in building a reaction database will know that atom mapping reagents/starting materials onto products is a very time-consuming and tedious process, that is often fraught with errors. So any method that automates this process is a significant step forward.
Here, we demonstrate that Transformer Neural Networks learn atom-mapping information between products and reactants without supervision or human labeling. Using the Transformer attention weights, we build a chemically agnostic, attention-guided reaction mapper and extract coherent chemical grammar from unannotated sets of reactions. Our method shows remarkable performance in terms of accuracy and speed, even for strongly imbalanced and chemically complex reactions with nontrivial atom-mapping. It provides the missing link between data-driven and rule-based approaches for numerous chemical reaction tasks.
They also supply a file containing Common patent reaction templates. This file contains the most common patent reaction templates (USPTO grants), including the year of the first appearance, the patent numbers, frequently used reagents, and the template count. The templates were extracted after applying RXNMapper to generate the atom-mapping.
A new version of DataWarrior has been released
v05.05.00: April 2021
- 3D-Structure alignment considering shape and pharmacophoric features (PheSA)
- Google Patent search and results in DataWarrior (keyword, structure, date, ...)
- Link to Spaya synthesis planning server
- Searchable and navigatable user manual
- Macro to retrieve and visualize world-wide Corona virus spreading
- Lots of new features, range filter animations, smarter labels, ...
The very popular bioinformatics tool MacVector 18.1 is now available to download. MacVector 18.1 is a Universal Binary application, which means it runs natively on both Apple Silicon M1 Macs and Intel Macs. MacVector 18.1 matches the “Big Sur” look and feel. …and for the first time in many, many years the MacVector icon has changed to match the square look of macOS Big Sur icons.
We ran some benchmarks to see how much faster MacVector now runs on an Apple Silicon MacBook Pro. We compared this against MacVector 18.0, which runs using Rosetta2 emulation. In some cases you can see that the native Apple Silcon MacVector 18.1 runs 200% faster than the emulated MacVector 18.0.
XQuartz 2.8.0 has been released for macOS 10.9 or later. I've been in touch with a couple of users and they report no issues so far. This is the first version with Apple Silicon support.
The XQuartz project is an open-source effort to develop a version of the X.Org X Window System that runs on OS X. Together with supporting libraries and applications, it forms the X11.app that Apple shipped with OS X versions 10.5 through 10.7.
Changes in 2.8.0
- Adds native support for Apple Silicon Macs.
- Removes support for versions of macOS older than 10.9
- Uses system libXplugin
- Removes build-time support for deprecated X11 libraries:
- Removes deprecated commands:
- Removes xpyb
- Removes older libpng
Full release notes are here https://www.xquartz.org/releases/XQuartz-2.8.0.html
A new version of ChemDoodle is available, this is a free update for all subscribers.
Whilst there are a few bug fixes and stability improvements the big news is the new feature enabling generation of a chemical structure from an image.
ChemDoodle now has the ability to recreate a chemical drawing from an image of a molecule, recovering the original chemical data. This function is performed using the File>Recover from Image... menu item. This function is different from inserting the image as inserting an image provides you with just a graphic, while recovering the chemical drawing allows you to regain access to the chemical information to use or edit further. We call this function Chemical Image Recovery (CIR). Some may also refer to this function as Optical Structure Recognition (OSR).
I can see this being a very popular feature.