[PDF] Design and Use of Relational Databases in Chemistry TJ O’Donnell

Relational databases and the relational model are far from new concepts, having first appeared 40 years ago. When we think of where data lives, we usually think of it residing in a relational database of some sort. However, there are domains like chemical informatics where the uptake of these databases is uneven – there are still many cheminformatics scientists who have never written a line of SQL, and still rely on antiquated models of large dataset handling and storage – like flat files in a unix system. This book aims to change that. From the first sentence of chapter 1: “The goal of this book is to convince you that relational databases are the best way to store, search, and even operate on chemical information.” By the time I had finished reading this book, I was indeed convinced. From the fundamentals of relational databases, to SQL, to IT considerations like client/web applications and how to store, search, and manipulate data, O’Donnell moves on to cover how chemistry is actually represented in such databases and what you can do with it. To make this practical he discusses how to implement an actual chemistry Relational Database Management System (RDBMS) in PostgreSQL, a popular open-source RDBMS. The book is a manageable size – 229 pages – which seems about right, of which 35 or so comprise 9 very helpful Appendices including lots of code snippets, which for me are the best way to learn. This is the kind of book I still keep out on my desk for reference to things like Table A.4 in the back. While this may seem too esoteric to some readers, it is to me an invaluable resource describing the SMARTS (a chemistry substructure specification language) implementation of one of the most popular sets of chemistry “keys” used in cheminformatics practice. Someone so inclined could learn much of the SMARTS language just by studying how these 166 keys are encoded by SMARTS strings. Readers will probably find numerous other things of this nature in the book that can be applied to their own work. O’Donnell has made a valuable contribution to the domain of chemical informatics, and I hope he convinces others of the power of relational databases in chemistry.