Question

I know the powerful MolDb5 http://merian.pch.univie.ac.at/~nhaider/cheminf/moldb5.html

• "Free" (open source at least, MySQL or Postgresql DBs)

• Algorithm and methods available.

• Underlying DB schema available

• Structure and functional group/fragments search, with a structure editor front-end -

• Accesible via web services

BTW, i'm talking about a "tool", not about a certain algorithm or search code!!! Are available other free tools like this?

Answer 1

Mychem offers two types of structure search:

• the first is based on fingerprint. You can use tanimoto scoring, as well as any other type of scoring using bit count (bit functions are provided by Mychem). You can also concatenate different types of fingerprint.
• the second is based on the use of smart, like obgrep. However, it's much more faster, as it uses binary objects.

These two methods are also provided by pgchem.

It would be also interesting to have an Open Dataset with the following requirements:

• a good molecular diversity;
• contains more than 50.000 molecules;
• contains racemic as well as chiral structures.

So, we could better compare the functions with the following criteria:

• accuracy
• speed
• scoring

I would suggest also to consider as important these parameters for choosing a good software:

• A good documentation
• A bug tracking system or a valid e-mail address to contact the author of the software.

Answer 2

Please also have a look at MyChem and PgChem.

Answer 3

I have been building a Structure database for Bioclipse. Right now it might be a bit of a proof of concept only but feel free to have a look. It uses a 100% Java database engine (HSQLDB) and it is not exactly super fast. I have tested it with the ChEBI complete (had to remove the bucky ball which CDK crashes on), that file is about 70 MB and contains about 13000 mols and that is about the amount of data that is currently reasonable to run in this very (as of now) unoptimized structure database.

The gui parts and the general design can most certainly be reused with one of the other more optimized structure databases as a base though. Some information about the project and a special build of Bioclipse with it can be found here

Answer 4

obgrep from OpenBabel.