ChemicBook

Large Language Models in Chemistry

Unless you’re living under a rock, you probably have heard about ChatGPT by now. The Internet has been buzzing with new technologies for the past few months. Tools that would have taken months, possibly years, to build are being developed just over a weekend of hacking. I truly believe that we’re living in one of the most exciting times in history.

A theoretical introduction to Unified Medical Language System (UMLS) for beginners

Let’s try to get a brief overview of what UMLS is, and then we’ll jump into more details. So, in simple terms, UMLS is a collection of many controlled biomedical vocabularies to provide cross-walk among them and to enable interaction between computer systems. To simplify it further, there are multiple biomedical vocabularies out there developed by different organizations. Each of them has come up with their own set of IDs assigned to each term. UMLS helps in mapping each of them to one common ID.

Predicting the products of a chemical reaction using graph-convolutional neural network (Part 2)

This is a continuation of a 2 part series in predicting the products of a chemical reaction using deep learning methods. Checkout the first part here.

In part 1, we:

1. First introduced the problem statement
2. Then we saw how a molecule is represented in a graph network

Predicting the products of a chemical reaction using graph-convolutional neural network (Part 1)

Reactant A and B reacts to form the products C and D. The chances are that you have encountered this type of chemical reaction in your school days. While on paper they may seem as simple as drawing a bunch of arrows, in the real world, even the simplest of those reactions, takes a long time, money, and resources to formulate. The reaction time can vary from the order of seconds to the order of months.

A beginner's guide for understanding Extended-Connectivity Fingerprints(ECFPs)

Extended-Connectivity Fingerprints(ECFPs) are a type of molecular fingerprint explicitly designed to capture molecular features relevant to molecular activity. They are among the most popular similarity search tools in drug discovery and they are effectively used in a wide variety of applications.