Home

SmilesDB

A SMILES-first molecular database

smilesdb.org

Motivation

SmilesDB is a product of need - while working on a research program at the Wolfram Emerging Leaders Program, I found that there were no readily accessible online SMILES databases.

There were, of course, databases with SMILES strings, but it was difficult to query these in large quantities (quickly).

Technical Details

SmilesDB runs on Render servers and is written on the Flask microframework. Web scraping was done in advance and used BeautifulSoup4 and Javascript snippets injected into downloaded HTML.

The API is documented and open source.

Challenges

The initial scraping was a challenge, as I sourced molecules from BindingDB - a site that uses dynamic loading for its tables. The ultimate solution involved javascript injected into the page.