[Dock-fans] 回覆︰ Dock-fans Digest, Vol 92, Issue 2

John Irwin jir322 at gmail.com
Thu Jan 12 10:44:32 PST 2012

Dear Dong

Thanks for your interest in a large custom ZINC subset. What you ask
for is a slight variation (and a subset) of ZINC fragment like:

ZINC Fragment like is MWT <= 250,  RB <= 5 and calculated LogP <= 3.5

You asked for

150 <= MWT <=250; calculated logP <= 2.5; RB <= 3;
plus H-donors <=3; H-acceptors <=4

How to proceed:  Please download ZINC fragment like and filter out the
molecules that don't qualify by your criteria.  Please understand that
we cannot support allowing everyone to download custom
multi-hundred-thousand or even millions subsets of ZINC.  But should
be easy for you to customize the standard subsets locally.

Separately, can you help me understand why you want these limits?  For
instance, are you worried about the solubility of 250 amu compounds?
We think the converse.  We used to use logP of 2.5, but we upped it to
3.5 because we were missing too many perfectly good compounds.  Recall
that calculated logPs typically have an uncertainty of +/- 1 log.

What are the H-donors and H-acceptor limits for? Why would you not
want a fragment with 5 H donors?  This rule excludes glucose, for
instance, as well as many other perfectly acceptable molecules, in our
view.  If you are worried about molecules being too soluble, you might
be better off with a lower limit on calculated logP, say -1 or so,
although that would exclude glutamate (logP of -4), which is a very
important ligand.

Why the rotatable bond limits?  Do you really want to exclude lysine
and arginine?  What is the purpose of excluding small flexible

Finally, what is the point of a lower limit on molecular weight?  You
only eliminate about 34K of 583K, less than 6%.   I admit to
considering a lower limit myself from time to time, but people keep
coming to me with projects where phenol (94 amu) or catechol (110 amu)
can be interesting, now that SPR and NMR are so widely used.  For most
projects I have seen, cutting at 150 amu does not help, and can make
you miss revealing things.

So, basically, I'm trying to encourage you to consider option 2 or even 3 again.

Good docking!


On Thu, Jan 12, 2012 at 12:33 AM, LIU Changdong <dongyi2827 at yahoo.com.hk> wrote:
> Dear John,
>   Thanks for your quick reply!
> Because of time limited, I prefer to the option 1 at first.(I'll try the
> option 2 at the same time).
> If it's ok, would you please help to make it from the lead-like library
> based on the below criteria?
> Molecular weight  150 <=  MW  <= 250
> xlogP<=2.5
> Rotatable bonds <=3
> Hydrogen donors <=3
> Hydrogen acceptors  <=4
>  Thanks again!
> Best
> Dong
> Message: 2
> Date: Wed, 11 Jan 2012 10:11:21 -0800
> From: John Irwin <jir322 at gmail.com>
> Subject: Re: [Dock-fans] create library for virtual screening from
>     ZINC
> To: LIU Changdong <dongyi2827 at yahoo.com.hk>
> Cc: "dock-fans at docking.org" <dock-fans at docking.org>
> Message-ID:
>     <CAHj+jShbMpCZUQXbNP_TvLBB=YHGuvXve4SxOHdAk0xqdgO4ww at mail.gmail.com>
> Content-Type: text/plain; charset=Big5
> Dear Dong
> Thanks for your question about preparing a library in ZINC.  ZINC
> enables you to get large pre-made libraries by properties and vendors,
> or small (<5000) libraries by arbitrary criteria. We cannot
> (currently) support large arbitrary criteria bigger than 5000, because
> of the load it would put on our public server.
> If you want libraries selected using arbitrary criteria bigger than
> 5000 molecules, you have three options:
> 1. ask me. If it is something that might be popular, we can make it for
> everyone
> 2. download a superset in ZINC (e.g. drug-like), and do the selecting
> on your local machine
> 3. realize that one of our pre-made subsets, while not *exactly* what
> you want, is close enough for your purposes.
> Good docking and good luck!
> John
> PS.  If you don't log in, you only get 500 molecules.  Use the "next"
> button (top left of the listing) to go to the next page.  If you log
> in (free registration!) you can get up to 5000 molecules per query.
> You only get 100 molecules per page, and you use the next button, or
> type the page number you want to jump to.
> PPS we are working on allowing arbitrary criteria for subsets of
> 50,000 molecules available to the general public, which we hope will
> be on line before the end of the year.  We are unlikely to support
> arbitrary subsets larger than 50,000,  but who knows.
> On Wed, Jan 11, 2012 at 12:30 AM, LIU Changdong <dongyi2827 at yahoo.com.hk>
> wrote:
>> Dear DOCK fans,
>>  I'm just preparing a library for virtual screening from ZINC.
>>  I'm using the search function on the ZINC website to search the compounds
>> which I need, my question is:
>> 1) after pressing "run query", there's only 1 page( for example 100 for 1
>> page), how can I know how many compounds I got based on my criteria.
>> 2) How can I download all the compounds which I need, every time I can
>> only
>> download  500 compounds in mol2 format.
>> Thanks in advance for any help!
>> Best
>> Dong
>> _______________________________________________
>> Dock-fans mailing list
>> Dock-fans at docking.org
>> http://mailman.docking.org/mailman/listinfo/dock-fans
> ------------------------------
> _______________________________________________
> Dock-fans mailing list
> Dock-fans at docking.org
> http://mailman.docking.org/mailman/listinfo/dock-fans
> End of Dock-fans Digest, Vol 92, Issue 2
> ****************************************
> _______________________________________________
> Dock-fans mailing list
> Dock-fans at docking.org
> http://mailman.docking.org/mailman/listinfo/dock-fans

More information about the Dock-fans mailing list