TopMap-SD

A Shape-Distribution-based molecular topology mapping algorithm.

The software provided here is to be described soon in detail under the putative title:

"A fast topological analysis algorithm for large-scale similarity evaluations for ligands and binding pockets"

Authors: Mohammad ElGamacy and Luc Van Meervelt.

Additional modification or custom updates requests are welcome and can be communicated to the following email address:

mohammad.elgamacy@vub.ac.be

Usage

TopMap hashes molecular topolgies into a series of moments. Molecular vectors records are generated in ./outfile.dat

• Input structural coordinates must be provided in mol2 format with the atomic partial charges assigned and preferrably exported/reformatted using OpenBabel. Essential records (@<TRIPOS): - MOLECULE - ATOM - BOND
• Input mol2 files can be input as gzipped (i.e. ".mol2.gz") files
• Output file contains the molecular descriptor vectors in text format (for portability reasons; UTF-8).
• Output data is flushed after each molecular calculations loop cycle, thus, the output file completeness can provide an indication of the computation progress.

Usage: topmap[.exe] mol2_input_file[.gz]

Example

An example scenario on a dataset from the DUD-E is presented below:

The example shows the main steps of structural fingerprinting (without any conformational sampling steps) and running a simple search using a random query structure.

This illustration was made for an Ubuntu system. Example mol2 files can be downloaded from any target dataset from the DUD-E (http://dude.docking.org/targets)

step 1:

download and extract the zipped package

unzip TopMap-SD-master

step 2:

fingerprinting actives:

./TopMap_v0.1linux actives_final.mol2

fingerprinting decoys:

./TopMap_v0.1linux decoys_final.mol2

contactenate output files:

cat actives_final.dat decoys_final.dat > all_fp.dat

prerequisite:

Numpy library must be installed (http://www.numpy.org/) for the next step

step 3:

calculating dissimilarity, sorting similar molecules to a query, and printing out the results to the stdout. This calculates the dissimlarity and prints out the names and dissimilarity scores of the top 50 similar hits using the first structure in the file (i.e. ligand nr. 1) as query: python dssmlrty_cmp_minima.py 1 50 all_fp.dat