TLDR: Difference between revisions

From DISI
Jump to navigation Jump to search
(22 intermediate revisions by 2 users not shown)
Line 8: Line 8:


== Current available modules ==
== Current available modules ==
=== Arthorbatch ===
* [[TLDR:arthorbatch|Arthorbatch]]
This modules allows user to by pass the 20,000 result limit in arthor.docking.org. It supports search by SMILES and by SMARTS.
* [[TLDR:bioisostere|Bioisostere]]
=== Bioisostere (formerly Analog) ===
* [[TLDR:bootstrap1|Bootstrap1]]
The purpose of the bioisostere app is to find analogs of your active compounds that you can use to explore SAR by catalog around your hits.
* [[TLDR:bootstrap2|Bootstrap2]]
 
* [[TLDR:dude-z|DUDE-Z]]
This app requires:  
* [[TLDR:extrema|Extrema]]
* a starting molecule (one at a time for now)
* [[TLDR:newbuild3d|Newbuild3d]]
* a tanimoto limit
* [[TLDR:strain|Strain]]
* mwt and logP limits
* [[TLDR:swbatch|Swbatch]]
* which bioisosteric replacements are acceptable (default all)
This app returns:
* 2D molecules within limits on molecular weight and logP (default 400 and 4.0 respectively) and tanimoto (default 0.5).
* Results are in three separate files.
** a. molecules you can buy according to ZINC
** b. molecules you can make or have made in one, two or three steps from commercially available building blocks
** c. molecules that are intriguingly similar but will require more synthetic effort.
 
The output of this app can be used by the following other modules in TLDR
* blaster, for docking.
* build 3d, to build docking libraries
* potentially covalent, depending on warhead
* cluster, to cluster the results
* dude, to generate decoys
* libanalysis, to get a detailed analysis of properties and targets
* report2d, to get a quick summary of physical and chemical properties
 
=== Bootstrap1 ===
Compare performance of 2 docking methods with p value.
=== Bootstrap2 ===
Get 95% confidence interval from bootstrapping.
=== DUDE-Z ===
Purpose: Assemble decoys for docking using the DUDEZ approach. See http://dudez.docking.org for sample data
 
This app requires:
* Ligands provided in smiles format (actives.smi)
* Current default parameters are shown in the following decoy_generation.in file:
 
    PROTONATE YES
    MWT 0 125
    LOGP 0 3.6
    RB 0 5
    HBA 0 4
    HBD 0 3
    CHARGE 0 0
    LIGAND TC RANGE 0.0 0.35
    MINIMUM DECOYS PER LIGAND 20
    DECOYS PER LIGAND 50
    MAXIMUM TC BETWEEN DECOYS 0.8
    TANIMOTO YES
 
See http://wiki.docking.org/index.php/Generating_decoys_(Reed%27s_way) for an explanation of what these parameters mean and how DUDE-Z works.
 
This app returns:
* List of decoys
 
=== Extrema ===
Generate an equivalent number of charged (-2, -1, 0, +1, +2) molecules in the molecular weight and cLogP property space of your input SMILES molecules. These charged molecules can be used as a decoy background for docking to identify the charge preference of your docking setup.
=== Newbuild3D ===
Prepare a 3D library for docking in several formats used by popular docking programs: sdf, db2, mol2, pdbqt, solv


== In Development ==
== In Development ==
Line 200: Line 149:


= Technical info =  
= Technical info =  
== Ben's info section ==
python environment is @ /nfs/soft/www/home/apps/tools18/envs/tldr-prod/bin/python3.9
tldr distribution packages are @ /nfs/soft/www/home/apps/tools18/dist
=== How to open source code & make a change ===
1. unpack most recent distribution in dist folder to your preferred directory
2. make changes to source code. this is left as an exercise for the reader
3. once done, re-pack source code into tarball. if you want to change the version name, change the name of the source code's root directory & dist package- for organizational purposes
4. copy the package back to the tldr dist directory- not strictly necessary, but again, for organizational purposes
5. with the tldr python environment, pip install the package you just created
* /nfs/soft/www/home/apps/tools18/envs/tldr-prod/bin/python3.9 -m pip install /nfs/soft/www/home/apps/tools18/dist/$YOUR_PACKAGE
* this will uninstall the previous version- I haven't tried doing this while the server is live, but it should probably(?) be fine
6. restart the tldr webserver on gimel2
* supervisorctl restart tldr
=== Making changes to modules & scripts ===
modules are located @ /nfs/ex7/tldr-modules
input files & parameters are defined by parameters.json within the module directory
modules are fairly straightforward bash scripts that operate in a designated working directory created by tldr when a job is submitted
job files are stored in /nfs/ex7/blaster/jobs/[0-9]
the specific 0-9 directory the job directory is located in is based on the last digit of the tldr-provided job_id- these directories are striped across multiple disks
within the job's working directory "results" is a special directory that will be shown to the user containing results
=== TLDR database ===
psql -h mem2 -p 5432 -d blaster -U blasteruser
can add new modules by adding to the job_types table
set module private/public by modifying job_type_status_fk in this table


== starting the server in single-threaded mode ==  
== starting the server in single-threaded mode ==  

Revision as of 20:27, 29 April 2024

TLDR (tldr.docking.org) is a web-based interface to molecular docking and related tools. The system currently consists of a dozen apps. We plan to support dozens, perhaps a hundred all told. Here we list the apps you can currently use, with usage notes.

This is also called "Add TLDR Module"

Current available modules

In Development

Blaster

The purpose of the blaster app is to prepare a receptor for docking, including some basic analysis.

This app requires:

  • A structure for your receptor protein provided as a pdb file
  • A binding site provided in one of the following ways:
 1) Supply ligand in binding site
 2) Provide binding site residues
 3) Use a program to identify all potential binding sites.  Choose which of the binding sites to test or test them all.

The app returns:

  • dockfiles, which may be used for large library docking
  • workfiles, which may be used for grid and sphere optimization.

The output of this app can be used by:

  • asdf
  • sef
  • sdfafd

Status: works. (equivalent of blastermaster.py)


Covalent

Purpose:

This app requires:

  • asdf

This app returns:

  • asdf

The output of this app can be used by:

  • adsfasdf

Status: Works with special cases only. Nearly ready to use. If interested, ask jji for assistance.

Cluster

Purpose:

This app requires:

  • asdf

This app returns:

  • asdf

The output of this app can be used by:

  • adsfasdf

Status: Works.


Libanalysis

Purpose:

This app requires:

  • asdf

This app returns:

  • asdf

The output of this app can be used by:

  • adsfasdf

Status: Not working yet

Reaction

Purpose:

This app requires:

  • asdf

This app returns:

  • asdf

The output of this app can be used by:

  • adsfasdf

Status: Works at a basic level, with minor caveats.

  • need to handle mwt and logP cutoff parametrically.
  • needs work to handle millions of molecules
  • needs work to connect to reagents, reactions and schemes

Report2d

Purpose:

This app requires:

  • asdf

This app returns:

  • asdf

The output of this app can be used by:

  • adsfasdf

Status: Not working yet

Shape

Purpose:

Search for similar shape molecules for ligands

This app requires:

  • asdf

This app returns:

  • asdf

The output of this app can be used by:

  • adsfasdf

Status: Not working yet

ZINCbatch

Purpose:

This app requires:

  • asdf

This app returns:

  • asdf

The output of this app can be used by:

  • adsfasdf

Status: Not working yet

Technical info

Ben's info section

python environment is @ /nfs/soft/www/home/apps/tools18/envs/tldr-prod/bin/python3.9

tldr distribution packages are @ /nfs/soft/www/home/apps/tools18/dist

How to open source code & make a change

1. unpack most recent distribution in dist folder to your preferred directory

2. make changes to source code. this is left as an exercise for the reader

3. once done, re-pack source code into tarball. if you want to change the version name, change the name of the source code's root directory & dist package- for organizational purposes

4. copy the package back to the tldr dist directory- not strictly necessary, but again, for organizational purposes

5. with the tldr python environment, pip install the package you just created

* /nfs/soft/www/home/apps/tools18/envs/tldr-prod/bin/python3.9 -m pip install /nfs/soft/www/home/apps/tools18/dist/$YOUR_PACKAGE

* this will uninstall the previous version- I haven't tried doing this while the server is live, but it should probably(?) be fine

6. restart the tldr webserver on gimel2

* supervisorctl restart tldr

Making changes to modules & scripts

modules are located @ /nfs/ex7/tldr-modules

input files & parameters are defined by parameters.json within the module directory

modules are fairly straightforward bash scripts that operate in a designated working directory created by tldr when a job is submitted

job files are stored in /nfs/ex7/blaster/jobs/[0-9]

the specific 0-9 directory the job directory is located in is based on the last digit of the tldr-provided job_id- these directories are striped across multiple disks

within the job's working directory "results" is a special directory that will be shown to the user containing results

TLDR database

psql -h mem2 -p 5432 -d blaster -U blasteruser

can add new modules by adding to the job_types table set module private/public by modifying job_type_status_fk in this table

starting the server in single-threaded mode

source /mnt/nfs/work/chinzo/Projects/BlasterX_supritha/venv/bin/activate
python code/DOCKBlaster/autoapp.py

How to add new module

See Add Tools18 module

Supported field types

For now, the model accepts "text_box", "check_box", "drop_down" , "radio_button",  and so on

If "type" is "text_box", it can contain a text or number with a min and a max range. If there is a min and a max range, then they have to be mentioned as "value_type": "number", "value_range": {"min_value": 0.1,"max_value": 0.99} as in parameters.json for cluster. If "type" is "text_box" and "value_type" is "text", then it is a normal text box with no range or validations. 6. Every input mentioned under the key "inputs" has a field called "file_name", which the name by which the input file uploaded/filled by the user gets stored in the file system at /nfs/ex7/blaster/jobs/JobID%10/Jobname_jobID folder.

7. Every job type has a "job_output" field, which currently stores an empty results.txt file which can be modified to do another action later. For now, the inputs uploaded, and the output file name specified by the user gets stored in the file system under the path that I mentioned in point 6.

Retrieved from "http://wiki.docking.org/index.php?title=TLDR&oldid=15747"