Generating extrema set

Updated 5/27/2020

Written by Reed Stein.

This is a wrapper script that performs everything below including JK's "gen_extrema.py" script. It requires a SMILES file and the number of molecules you would like from each molecular weight/cLogP tranche. Make sure to source python3

   source /nfs/soft/python/current/env.csh

Then run:

   python ~rstein/zzz.scripts/TLDR_DUDE/0001_db2_map_tranche_collect_db2_gz.py -s {SMILES_FILES}

You can also specify number of molecules to return. The default is 500.

   python ~rstein/zzz.scripts/TLDR_DUDE/0001_db2_map_tranche_collect_db2_gz.py -s {SMILES_FILES} -n 1000

This script will find the interquartile range of molecular weight/cLogP properties of your input SMILES ligands. Then it will retrieve db2.gz files for all -2, -1, 0, +1, +2 charges for each MWT/cLogP tranche.

Once this finishes running, as below:

  cat *_charge_tranches.list > extrema_set.list

which is the split_database_index you would then use for docking.

Written by Jiankun Lyu, 2019/10/12

The main purpose of the extrema set is to test the charge preference of your docking setup and to make sure that you don't over-optimize your docking setup with property-matched(charged-matched) decoys generated by DUDE. This is also a sanity check of your docking setup, put it on your checklist!

extrema_set_gen------- working 
                              |
                              |------ ZINC-downloader-3D-minu2.database_index
                              | 
                              |------ ZINC-downloader-3D-minu1.database_index
                              | 
                              |------ ZINC-downloader-3D-neutral.database_index
                              | 
                              |------ ZINC-downloader-3D-plus1.database_index
                              | 
                              |------ ZINC-downloader-3D-plus2.database_index

1) Make those directories above.

mkdir extrema_set_gen
cd extrema_set_gen
mkdir working

2) Download databases index from ZINC with different charge types

2.1) Go to ZINC http://zinc15.docking.org/tranches/home/#

2.2) Choose the tranches you want to generate extrema set for testing the charge preference. The goldilocks set has been chosen here as an example.

Choose the -2 charged tranches for your extrema set

Choose the -1 charged tranches for your extrema set

Choose the neutral tranches for your extrema set

Choose the +1 charged tranches for your extrema set

Choose the +2 charged tranches for your extrema set

2.3) download the databases index file for each charge type

download the databases index file for each charge type

2.4) download the files above and save it as ZINC-downloader-3D-(charge-type).database_index, then upload the file to the working directory. In the working directory, you are supposed to have 5 files with names: ZINC-downloader-3D-minu2.database_index, ZINC-downloader-3D-minu1.database_index, ZINC-downloader-3D-neutral.database_index, ZINC-downloader-3D-plus1.database_index and ZINC-downloader-3D-plus2.database_index.

3) Run extrema set generation on 5 different charge types

python /mnt/nfs/ex5/work/jklyu/sigma2/gen_extrema/script/gen_extrema.py
First input: the DB index from ZINC15
Second input: the prefix of the ligand charge
Third input: the lower bound of number of molecules for each tranche

python /mnt/nfs/ex5/work/jklyu/sigma2/gen_extrema/script/gen_extrema.py ZINC-downloader-3D-plus2.database_index 'plus2' 100 > log_plus2 &
python /mnt/nfs/ex5/work/jklyu/sigma2/gen_extrema/script/gen_extrema.py ZINC-downloader-3D-plus1.database_index 'plus1' 100 > log_plus1 &
python /mnt/nfs/ex5/work/jklyu/sigma2/gen_extrema/script/gen_extrema.py ZINC-downloader-3D-neutral.database_index '0' 100 > log_0 &
python /mnt/nfs/ex5/work/jklyu/sigma2/gen_extrema/script/gen_extrema.py ZINC-downloader-3D-minus1.database_index 'minus1' 100 > log_minus1 &
python /mnt/nfs/ex5/work/jklyu/sigma2/gen_extrema/script/gen_extrema.py ZINC-downloader-3D-minus2.database_index 'minus2' 100 > log_minus2 &

4) Output files

4.1) (charge-type)_tranche_summary.txt. The file contains how many molecules has been selected from each tranche. The section below is an example:

EF 430
DF 337
ED 293
DD 181
DE 120
CF 112
CE 131
CD 272
EE 118
1994

4.2) (charge-type)_charge_tranches.list. The file contains all the db2 indexes that have been selected from the extrema generation.

5) Combine all the db2 indexes generated by the script.

cat *_charge_tranches.list > extrema_set.list

6) Use the extrema_set.list as .sdi to set up your docking screen then run it.