Hi Martin

I believe that the main algorithm for the FixBondOrdersTool is actually
deterministic. However, as a prelude to sorting out the bond orders, the
rings which need fixing are first recorded thus (in the function
'removeExtraRings')

SSSRFinder arf = new SSSRFinder(m);
IRingSet rs = arf.findSSSR();

...code to remove non-aromatic rings etc...

return rs;

When I was testing this out, I noted that the IRingSet was differently
ordered on running multiple times and indeed the javadoc for findSSSR()
explicitly states that "the returned set is not uniquely defined." Most of
the other ring-finding routines in SSSRFinder are uniquely defined
according to the javadoc so perhaps it would be possible to use one of
these instead (I don't know enough about their behavior to know whether
this would be possible)? If this approach does work it's a one line fix.

Best wishes

Kevin

On Sun, Sep 28, 2014 at 3:36 PM, Martin Guetlein <
[email protected]> wrote:

> Hi Kevin,
>
> apparently you are the author of this very useful tool. Do you have
> any idea on how to get rid of the randomness (e.g., by setting a
> random seed)?
>
> Kind regards,
>
> Martin
>
>
>
> ---------- Forwarded message ----------
> From: Martin Guetlein <[email protected]>
> Date: 26 September 2014 15:49
> Subject: FixBondOrdersTool not determenestic
> To: cdkuser <[email protected]>
>
>
> Hi,
>
> At some point, FixBondOrdersTool apparently makes random choices and
> this causes my SDF export to produce different results with the same
> input on different runs.
> Can this be circumvented somehow?
>
> Thanks and kind regards,
> Martin
>
>
> --
> Dipl-Inf. Martin Gütlein
> Phone:
> +49 (0)761 203 8442 (office)
> +49 (0)177 623 9499 (mobile)
> Email:
> [email protected]
>
>
> --
> Dipl-Inf. Martin Gütlein
> Phone:
> +49 (0)761 203 8442 (office)
> +49 (0)177 623 9499 (mobile)
> Email:
> [email protected]
>
------------------------------------------------------------------------------
Slashdot TV.  Videos for Nerds.  Stuff that Matters.
http://pubads.g.doubleclick.net/gampad/clk?id=160591471&iu=/4140/ostg.clktrk
_______________________________________________
Cdk-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/cdk-user

Reply via email to