ChemTattoo 3D

ChemTattoo 3D is a freeware tool that analyzes a set of shape-aligned 3D conformers to find a set of common 2D patterns (defined by SMARTS) -- the 'modal' features for the dataset. After identifying the features ChemTattoo 3D creates a web-based depiction of the conformers that shows how well the modal features overlap in space.

An outline of the algorithm is as follows:


for each conformer:

for each modal feature type:

find the best overlap between any instance of that feature type in the target and in the conformer as before.

record the number of times each instance of a feature type, in the target, was the "best overlap" instance.

Then, after processing all of the conformers:

for each modal feature type, m:

find the instance of that feature type, in the target, which was most often the best match for a conformer.

highlight that as the target's match for m.


When running a ChemTattoo 3D analysis you can specify one of the input conformers as the target. If you do so, the algorithm determines the modal features of the target and then compares the features as well as their 3D distance to the remaining conformers.SMARTS patterns are defined as "Hbond Donor", "Hbond Acceptor", 6 membered ring, uncharged, etc. These patterns can be used as the default or you can specify your own SMARTS patterns for the algorithm to use.

If a target is not specified, the algorithm will choose one of the conformers as the target.\'a0 The chosen target is the smallest conformer which has the most features in common with the dataset.

ChemTattoo 3D expects an input SD file containing shape aligned conformers. The Mesa Analytics & Computing, Inc. shape fingerprinter produces shape aligned conformers.

Runtime dependencies:

'95 Python 2.6.x

'95 OpenBabel 2.2.3 with Python wrappers (openbabel and pybel)

On Ubuntu you will also need the libssl-dev package (0.9.8o-1ubuntu4.1).

To view results you'll need a Java-enabled web browser. If you're running on Linux you may need to install a Java runtime.


Example command line

python 0.9 -t 1

run ct3d on at a threshold of 0.9 using the first conformer in the dataset as the target output files go into the current directory.

python --help for options

Back to the Mesa website.