Indigo reaction mapping incomplete?

[Ernst-Georg Schmid]

Hello,

I am fairly new to reactions, so please bear with me if I'm completely wrong on this topic. :-)

I have used the following code to try Indigo's reaction automapper with 1.0.0 and 1.1-beta5:

#include <stdio.h>

#include <string.h>

#include <stdlib.h>

#include "mingw/indigo.h"

int main(int argc, char *argv[]) {

int handle;

qword session = indigoAllocSessionId();

indigoSetSessionId(session);

printf("%s\n",indigoVersion());

handle = indigoLoadReactionFromFile("c:/temp/testrxn.rxn");

indigoUnfoldHydrogens(handle);

printf("1:%s\n",indigoGetLastError());

indigoAutomap(handle,"discard");

printf("2:%s\n",indigoGetLastError());

indigoSaveRxnfileToFile(handle,"c:/temp/testrxn_mapped.rxn");

printf("3:%s\n",indigoGetLastError());

    indigoFree(handle);

indigoReleaseSessionId(session);

return 0;

}

And now I wonder why the C#N from the hydrocyanic acid is automapped, but not the C=O from the the methanal?

The Infochem reaction mapper maps them all (see tesrxn_mapped_infochem.rxn) and that seems the natural thing to do to me....

best regards,

Ernst-Georg

[Savelyev Alexander]

The given issue is well known and is under development now. Current
AAM engine can not handle with small MCS structures(< 2 atoms)

[Daniel Lowe]

Is this likely to make it into the next release?
On Nov 7, 8:10 am, Savelyev Alexander <asavel...@ggasoftware.com>
wrote:

[Savelyev Alexander]

The specified issue seems to be very simple, but I am busy with other
project at the moment. Therefore, I think, the problem will be
resolved in one-two weeks.

[Ernst-Georg Schmid]

I think I found another one:

If you look at the terminal C at the end of the double bond starting with atom 17 in educt A, it doesen't get a number at all. This propagates to the product side and the mapping numbers on the lower left ring change places with the upper ring. Besides just looking wrong, this breaks my R-Matrix. ;-)

Again, please see the infochem mapping for comparison.

best regards,

Ernst-Georg

[Ernst-Georg Schmid]

> on the lower left ring change places

That should be: on the lower right ring

[Savelyev Alexander]

Thank you for your suggestions. You are absolutely right about the
ring changes. I am going to fix the issues in the nearest future. I
think, the algorithm should consider all MCS solutions (it is not true
now) to select the most appropriate one.