Decoys: Difference between revisions

From DISI
Jump to navigation Jump to search
No edit summary
 
No edit summary
 
(16 intermediate revisions by 6 users not shown)
Line 1: Line 1:
Decoys are...
[[Decoy Theory |  Decoys are...]]


Decoys are important for judging the performance of [[molecular docking]] algorithms.  
Decoys are important for judging the performance of [[molecular docking]] algorithms.  


The [[DUD database]] is a free directory of useful decoys for [[virtual screening]].
If you want decoys for a molecule in ZINC, say 556, use
http://zinc15.docking.org/substances/ZINC000000000556/decoys/3D.sdf


== Scholarly Articles on Decoys ==
If you want decoys in 2D:
http://zinc15.docking.org/substances/ZINC000019632927/decoys/2D/


Alan
http://zinc15.docking.org/apps/mol/decoys?for=CN1CCN(CC(=O)N2c3ccccc3C(=O)Nc3cccnc32)CC1


== People working on Decoys ==  
== Examples ==


* Alan
=== Getting decoys for Aspirin (ZINC000000000053) ===
* Michael
    2D Decoys (need to generate DB2 files and compute charge yourself)
    Visualize:  <http://zinc15.docking.org/substances/ZINC000000000053/decoys/2D?>
    Download: <http://zinc15.docking.org/substances/ZINC000000000053/decoys/2D.smi?count=100>
 
    3D Decoys (Can DOCK directly)
    Visualize: <http://zinc15.docking.org/substances/ZINC000000000053/decoys/3D>
    Download (SMILES & Explicit Charge): <http://zinc15.docking.org/substances/ZINC000000000053/decoys/3D.smi?net_charge=-1&count=100>
    Download (DB2):  <http://zinc15.docking.org/substances/ZINC000000000053/decoys/3D.db2.gz?count=100>
 
Allowed args:
* count: How many
* resolve: Look up zinc_ids if possible
* unique: Only return unique decoys (if you POST with "for" as a file instead)
 
You could also do decoys "by hand" to have more control:
http://zinc15.docking.org/substances/?~ecfp4_fp-unsorted_tanimoto,.2=zinc55&mwt-between=270,330&logp-between=1.8,2.4&purchasability=for-sale
 
=== Getting Decoys for a SMILES (e.g. not in ZINC) ===
Getting Decoys for a SMILES string can be written as a general ZINC query on either the **substances** (2D) resource or the **protomers** resource (3D). Constraints can be added for further refinement.
Currently you need to have a smiles string (e.g. `CN(C)C(=O)c1ccc(O)cc1`) as well as the physical properties you of that compound (i.e. molecular weight, logp, etc.).  This requirement will soon be removed and replaced with a simple form.
 
Molecular Properties (Criteria) → URL Parameters:
 
    SMILES: CN(C)C(=O)c1ccc(O)cc1 (ECFP4-tanimoto < 0.25) → ~substance.ecfp4_fp-unsorted_tanimoto-25=CN(C)C(=O)c1ccc(O)cc1
    MWT: 165.2 (+/- 10%) → mwt-between=148.67,181.72
    LogP: 1.1 (+/- 0.30) → logp-between=0.8,1.4
    Donors: 1 (==) → hbd=1
    Acceptors: 2 (==) → hba=2
    Charge: 0 (==) → net_charge=0
    Rotatable Bonds: 1 (==)  → rb=1
 
As a ZINC query these constraints would be:
   
    <http://zinc15.docking.org/protomers/?~substance.ecfp4_fp-unsorted_tanimoto-25=CN(C)C(=O)c1ccc(O)cc1&mwt-between=148.67,181.72&logp-between=0.8,1.4&hbd=1&hba=2&net_charge=0&rb=1>
 
This query can be made faster by requesting only 50 and adding a few additional execution rules: count=50&parallelize=no&distinct=no
 
    <http://zinc15.docking.org/protomers/?~substance.ecfp4_fp-unsorted_tanimoto-25=CN(C)C(=O)c1ccc(O)cc1&mwt-between=148.67,181.72&logp-between=0.8,1.4&hbd=1&hba=2&net_charge=0&rb=1&count=50parallelize=no&distinct=no>
 
[[DUDE]] is a free directory of useful decoys for [[virtual screening]].
 
For more information on preparation see
 
[[Automated_Database_Preparation#Automatic_Decoy_Generation]]
 
 
[[Category:Jargon]]
[[Category:DOCK:Scoring problem]]
[[Category:DOCK:Sampling problem]]

Latest revision as of 18:38, 15 September 2015

Decoys are...

Decoys are important for judging the performance of molecular docking algorithms.

If you want decoys for a molecule in ZINC, say 556, use

http://zinc15.docking.org/substances/ZINC000000000556/decoys/3D.sdf

If you want decoys in 2D:

http://zinc15.docking.org/substances/ZINC000019632927/decoys/2D/
http://zinc15.docking.org/apps/mol/decoys?for=CN1CCN(CC(=O)N2c3ccccc3C(=O)Nc3cccnc32)CC1

Examples

Getting decoys for Aspirin (ZINC000000000053)

   2D Decoys (need to generate DB2 files and compute charge yourself)
   Visualize:  <http://zinc15.docking.org/substances/ZINC000000000053/decoys/2D?>
   Download: <http://zinc15.docking.org/substances/ZINC000000000053/decoys/2D.smi?count=100>
   3D Decoys (Can DOCK directly)
   Visualize: <http://zinc15.docking.org/substances/ZINC000000000053/decoys/3D>
   Download (SMILES & Explicit Charge): <http://zinc15.docking.org/substances/ZINC000000000053/decoys/3D.smi?net_charge=-1&count=100>
   Download (DB2):  <http://zinc15.docking.org/substances/ZINC000000000053/decoys/3D.db2.gz?count=100>

Allowed args:

  • count: How many
  • resolve: Look up zinc_ids if possible
  • unique: Only return unique decoys (if you POST with "for" as a file instead)

You could also do decoys "by hand" to have more control:

http://zinc15.docking.org/substances/?~ecfp4_fp-unsorted_tanimoto,.2=zinc55&mwt-between=270,330&logp-between=1.8,2.4&purchasability=for-sale

Getting Decoys for a SMILES (e.g. not in ZINC)

Getting Decoys for a SMILES string can be written as a general ZINC query on either the **substances** (2D) resource or the **protomers** resource (3D). Constraints can be added for further refinement. Currently you need to have a smiles string (e.g. `CN(C)C(=O)c1ccc(O)cc1`) as well as the physical properties you of that compound (i.e. molecular weight, logp, etc.). This requirement will soon be removed and replaced with a simple form.

Molecular Properties (Criteria) → URL Parameters:

   SMILES: CN(C)C(=O)c1ccc(O)cc1 (ECFP4-tanimoto < 0.25) → ~substance.ecfp4_fp-unsorted_tanimoto-25=CN(C)C(=O)c1ccc(O)cc1
   MWT: 165.2 (+/- 10%) → mwt-between=148.67,181.72
   LogP: 1.1 (+/- 0.30) → logp-between=0.8,1.4
   Donors: 1 (==) → hbd=1
   Acceptors: 2 (==) → hba=2
   Charge: 0 (==) → net_charge=0
   Rotatable Bonds: 1 (==)  → rb=1

As a ZINC query these constraints would be:

   <http://zinc15.docking.org/protomers/?~substance.ecfp4_fp-unsorted_tanimoto-25=CN(C)C(=O)c1ccc(O)cc1&mwt-between=148.67,181.72&logp-between=0.8,1.4&hbd=1&hba=2&net_charge=0&rb=1>

This query can be made faster by requesting only 50 and adding a few additional execution rules: count=50&parallelize=no&distinct=no

   <http://zinc15.docking.org/protomers/?~substance.ecfp4_fp-unsorted_tanimoto-25=CN(C)C(=O)c1ccc(O)cc1&mwt-between=148.67,181.72&logp-between=0.8,1.4&hbd=1&hba=2&net_charge=0&rb=1&count=50parallelize=no&distinct=no>

DUDE is a free directory of useful decoys for virtual screening.

For more information on preparation see

Automated_Database_Preparation#Automatic_Decoy_Generation