Drug Design and Discovery

L

Ligand

New Member
Joined
Mar 1, 2009
Messages
8
Reaction score
0
Since this is an IT website with a lot of knowledgeable people when it comes to software and hardware, and there are sure to be people that are interested in science, I thought I might try my luck and see if anyone with a little spare time and extra CPU speed to spare might be interested in designing or finding new biologically active compounds that might be relevant to a particular disease. For example AIDS, malaria, cancer, TB etc. It is relatively easy and all you need is the software.
Two snags:
1) You need to belong to an academic institution (say you’re a student in UCT).
2) You need to know a little about Linux


Two programs are freely available if you belong to an academic institution.
Just apply for the software.
DOCK 6.2 and Autodock.

With these programs you can see how a molecule binds to a protein, and if a ligand (that is just a term used for a compound that binds to a protein) binds with a good energy, you have a possible winner as a medicinally relevant compound.

There are libraries of compounds that contain virtually millions of ligands that have not been tested. You can run this software over weeks and test each compound and see how it binds to a specific site in a protein.

It takes about 5-10 minutes to see how a single compound binds to a protein on a 2.8 GHz CPU. If you have a quad core lying around, you can run the software on 3 of the "CPUs" and happily go on with the rest of your work/gaming etc.

Say you find a molecule that looks good (good position, good binding energy etc), what then? Well some of these compounds are commercially available, some can be synthesized. If you find one, there are institutions that will look into it and perhaps even do a quick biological test to see if it is active. Who knows, you can contribute to science and perhaps even discover a cure for AIDS, TB, cancer, malaria, diabetes etc.

Give a shout if you are interested ;).
 
I see HIV is a hot topic here. Why not try and design or find novel compounds that can inhibit or suppress HIV activity.

Take the HIV protease 1 protein. It plays an essential part in the life-cycle of the virus by cleaving (chopping up) proteins in the right place in order to generate mature protein components needed to assemble the virus. Without this protein, production of infectious HIV viral structures grind to a halt.

The protein is a target for drug discovery/design. How? Here is a visual of how the protein works and how it can be targeted.
3 compounds are used to inhibit the activity of this protein, namely:
L-735,524
ABT-538
Nelfinavir mesylate (Viracept)

However, due to associated toxicity, poor pharmacokinetic properties (bioavailability and metabolic degradation), and resistance that has developed to these drugs, new more effective drugs to inhibit this protein is needed.

Luckily there are several protein structures available and a massive library of compounds that have not been fully explored for novel inhibbitors.

The ZINC ligand database contains ligands that are already prepared for ligand docking and all that is needed are people willing to help and look for novel inhibitors of proteins that are medicinally relevant (like HIV protease).

Here is an example of the what the HIV protease looks like and Viracept bound to it.
picture.php

Viracept bound to HIV protease​

Current drugs inhibit this enzyme at about 22-180 nanomolar. The aim is to get compounds that can inhibit the activity of this protein in the single digit nanomolar range with good bioavalaibility and less cytotoxicity.

So, anybody interested in giving HIV a crack? It really is simple.
 
Last edited:
Very interesting. Pity I don't belong to an academic institution (and am probably too clueless to actually operate the software).
 
BlueCollar said:
Very interesting. Pity I don't belong to an academic institution (and am probably too clueless to actually operate the software).
Click to expand...
Ah man, that is a pity.

It really is easy. Once you have the program, all you need is a protein in 3D and a list of ligands (molecules) you think might be active. Basic Linux skills to run simple scripts are needed and I will guide you step by step.

This is basically the future of drug development.
1) Find molecules through docking.
2) Synthesize or buy them
3) Test their effect on gene expression through microarrays and other high-throughput screening methods in order to determine the exact mechinism of action.
4) Proceed to in vivo trials

With this technology the time it takes to develop drugs can be dramatically reduced.

Perhaps it can be relevant to hardware and software discussions as well? Sort of like a benchmark?
See whose computer can dock the most ligands in the shortest time.
Maybe have a CPU and RAM benchmark competition (because these are CPU and RAM intensive programs) and in the process discover new drugable compounds. The possibilities are endless.

Someone here interested?
 
Ligand said:
Say you find a molecule that looks good (good position, good binding energy etc), what then? Well some of these compounds are commercially available, some can be synthesized. If you find one, there are institutions that will look into it and perhaps even do a quick biological test to see if it is active. Who knows, you can contribute to science and perhaps even discover a cure for AIDS, TB, cancer, malaria, diabetes etc.
Click to expand...

Is this out of altruism, or who gets the big $$ when you hit the jackpot?
 
is this way accurate though? i mean, can a pc really substitute the real microbiology aspect ?

it looks cool and all but has it worked in the past ?
 
rwenzori said:
Is this out of altruism, or who gets the big $$ when you hit the jackpot?
Click to expand...
Hehe, I wouldn't be sharing this if I wanted to make big bucks :p.

Keeper said:
is this way accurate though? i mean, can a pc really substitute the real microbiology aspect ?

it looks cool and all but has it worked in the past ?
Click to expand...
As far as accuracy goes, it is not 100% accurate for all proteins. It is best used as a screening method when faced with a massive database. After a few hits are generated, the best way to test them would be to do biological assays. I don't think the PC is going to replace biological assays, though it can aid in searching massive databases or test designed ligands for possible leads.

Is it successful?
A recent study has demonstrated that this approach is feasible. Check out this study:
Chemical Probes that Competitively and Selectively Inhibit Stat3 Activation
Reported here.
There have been many other successes a well.

dappled said:
I'm sure a lot are interested but don't meet either requirement.
Why no windows version for us non-academia plebs?
Click to expand...
You can use the cygwin program for these programs, however running scripts on them can be a problem and it is slower than Linux. Something to do with cygwin and cache usage I think. Limited to academia? Well I think they are doing people in academia a favour by allowing them to use it for free. Companies are of course welcome to use it at a fee. If it is for purely academical purposes (like now, when you sign up for it), it is unfortunately limited to academia.
 
There is a new, and reasonably accurate program that everyone can use to dock ligands into proteins to see how they bind, and it makes use of multi-core systems. Takes about 1-2 minutes per ligand on a quad-core CPU., and it is reasonably accurate (RMSD < 2 for > 80 % of my proteins so far).
AutoDock Vina.
If anyone is interested, I can make a tutorial to make use of the program to test 1000s of compounds on your PC...
Give a shout :).
 
Last edited:
Ok, AutodockVina has officially been released. It looks good, it is free and it is accurate.

Here is a little tutorial (for Linux users) to dock +-100 ligands into the HIV reverse transcriptase protein and to see how it goes. Maybe a new and novel inhibitor capable of stopping HIV can be discovered?

  1. Make the following directory.
    In a terminal, do the following:
    mkdir ~/docking/MyBB
    cd ~/docking/MyBB

  2. Download AutodockVina (Linux version)
    Install it as described here.

  3. Download MGLTools 1.5.4
     (Whichever version of Linux suits you)
    Install it as described here. Note this is for the 1.5.2 version, but it should work for the 1.5.4 version.

  4. For this docking procedure, the HIV-reverse transcriptase (from here) is going to be used. Getting ligands to bind to this protein can lead to potential HIV drugs.
    There is a whole process to prepare the file. But here is he final version of the receptor protein file to be used.
    Link (Access it with any gmail account)
    When the above link open, do the following:
    Click "File"
    Click "Export"
    Click ".txt Sheet Only"
    And save the "receptor.tsv" in your MyBB folder.
    Rename the "receptor.tsv" file to "receptor.pdbqt". (This will allow AutodockVina to read the file)

    If there is another protein you would like to probe, e.g. malaria or TB related protein, you are welcome to give a few suggestions.

  5. Download the ligands to be docked from here.
    Click "File"
    Click "Export"
    Click ".txt Sheet Only"
    And save the ligands.tsv in your MyBB folder.
    Rename the "ligands.tsv" file to "ligand.mol2". (This will allow AutodockVina to read the file)

  6. Download the docking protocol here.
    Click "File"
    Click "Export"
    Click ".txt Sheet Only"
    And save the "vina_docker_mybb.tsv" in your MyBB folder.
    Rename the "vina_docker_mybb.tsv" file to "vina_docker_mybb".

  7. Download the docking parameters here.
    Click "File"
    Click "Download file as"
    Click "Text"
    And save the "MyBB_conf_VS.txt" in your MyBB folder.

  8. Check that you have the following files in your ~/docking/MyBB folder:
    MyBB_conf_VS.txt
    vina_docker_mybb
    ligands.mol2
    receptor.pdbqt

  9. Open MyBB_conf_VS.txt.
    Replace "INSERTHOME" with your home directory.
    Note to users with multicore CPUs:
    You can change the "cpu = 1" value to the amount of CPUs you wish to use for the application.

  10. Now run the docking simulation by:
    A) Opening a new terminal
    B) cd ~/docking/MyBB
    C) chmod +x vina_docker_mybb

    And finally to run the docking:
    D) ./vina_docker_mybb

Let's see who can get the fastest time to dock all those ligands and the lowest scoring (meaning best) ligand.
Happy docking! :)
 
You must log in or register to reply here.
Back
Top
Sign up to the MyBroadband newsletter
X