myChEMBL on Mac OS X

Over at the ChEMBL blog, Michal Nowotka recently wrote about setting up myChEMBL without using a virtual machine. Why do this? The main benefit is that it’s a nice way to quickly set up your own machine with a local copy of the ChEMBL database along with the associated chemistry toolkits, database software, web services and learning materials.

As it was originally designed to be used within a virtual machine, the original myChEMBL bootstrap.sh script was really only written with Ubuntu in mind. But with a little bit of work I put together a modified bootstrap.sh script to set up most of the myChEMBL functionality on Mac OS X.

As this is intended to be run on an existing system instead of a fresh new virtual machine, there are some major differences in the methods used:

  • Instead of overwriting configuration files completely (e.g. for Apache and PostgreSQL), this script makes the necessary changes to existing files in an effort to preserve any customisations you might already have.
  • Instead of installing the myChEMBL launchpad and web app at the root level of the web server, files are installed inside the current user folder and are therefore accessible from http://localhost/~username/mychembl.
  • For safety, the Apache web server, PostgreSQL server and iPython notebook server are all configured to only listen to localhost and ignore requests from elsewhere by default.
  • As we are targeting OS X instead of Ubuntu, the homebrew package manager is used instead of apt-get.

The script hasn’t been widely tested, so please take a minute to read through and understand what it does before running it. It is going to take a long while, particularly to download the ChEMBL database and generate the relevant fingerprints and indexes.

You can find all the code on Github as mychembl-osx, a forked version of the original myChEMBL repository.