Scripted testing for aggregators

From DISI
Revision as of 21:18, 15 December 2016 by Frodo (talk | contribs) (asdf)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

To script testing for aggregators, you have two options, aggregator advisor and ZINC15.


Aggregator advisor

Prepare your molecules to be tested. Here we test the first 100 FDA approved compounds from ZINC

wget zinc15.docking.org/substances/subsets/fda.smi

The file looks like this

$ grep -n  ZINC000000020237 fda.smi  
32:CCN(CC)CC(=O)Nc1c(C)cccc1C ZINC000000020237
  • Run the program as
python simi.py  similar fda.smi > output
  • When there are matches to aggregators, each line contains :
    • original smiles
    • matched aggregator
    • tanimoto similarity (in the paper we suggest 0.85 cutoff, but we are more generous in the calculation here to allow you to make the cut
    • original identifier from the input file
  • e.g. the first line of output is as follows
$ python simi.py  similar fda.smi
CCN(CC)CC(=O)Nc1c(C)cccc1C	Cc1cccc(C)c1NC(=O)CNc1ccccc1	 0.87	ZINC000000020237


ZINC15

We now have a bulk processing facility, the right hand side of http://zinc15.docking.org/patterns/home/. It supports both PAINS and Aggregator similarity searches.

To use it, paste SMILES, one per line, or upload a file, up to 1000 per transaction. A whitespace-separated identifier after the SMILES is optional. It can be used with curl programmatically if you prefer.

For instance, if you enter: CCOc1ccc([C@@H]2[C@H]3CCCC=C3C(C#N)=C(N)C2(C#N)C#N)cc1 some_identifier

and click on "Check" you will see it matches: cyano_ene_amine_A(56)

If you enter: C1=CC(=C(C=C1C2=C(C(=O)C3=C(C=C(C=C3O2)O)O)O)O)OC blah_blah_blah

and click on "Check" you will see it is similar to an aggregator, and we have a button that allows you to run it again in the Aggregator Advisor as published recently in J Med Chem. We've used ECFP4 Tc 0.6 instead of axonpath 0.85 as in the paper, so it is an approximation. 0.55 ECPF4 might be better, we're still testing.


There is a way to script this. It is similar to but not exactly the same as the aggregator advisor. If you want the scripted version, please write me.

curl http://zincfifteen.docking.org/patterns/apps/checker.txt -F upload=@smiles_file -F pains=y -F aggregators=y

please do not over use this feature. Slow! works best with < 100 smiles per transactions

-- John