Alignment in Marvin* 5.4

When I generate an alignment of a set of molecules drawn in marvin/sketch and save the results as a plain SD-formatted file, I end up with a giant single molecule.  Is there any easy way to get the molecules written out as distinct molecules in a common coordinate system?  I'd like to try using Marvin to prepare alignments for 3D-QSAR methods, but having to manually transform the resultant single-molecule output back to a multi-molecule file is a major impediment.

 

Thanks,

Andy

afant wrote:

When I generate an alignment of a set of molecules drawn in marvin/sketch and save the results as a plain SD-formatted file, I end up with a giant single molecule.  Is there any easy way to get the molecules written out as distinct molecules in a common coordinate system?  I'd like to try using Marvin to prepare alignments for 3D-QSAR methods, but having to manually transform the resultant single-molecule output back to a multi-molecule file is a major impediment.   Thanks, Andy

Hi Andy,

There is a tool specialized for alignment that is not part of the Marvin suite yet. It needs java installed into your browser. This tool exactly does that you need.

If you have any difficulties in usage or any ideas for new features feel free to contact me here.

You can find it here:

http://prefect.chemaxon.hu/shared/akalaszi/alignment/

Adrian

Thanks Adrian.  I'm playing around with it a little now that I have some spare time, and I think I have the hang of it.  I'm going to have to find a way to subset the molecules to be aligned to fall within the limits of a non-templated search. 

 

Thanks again,

Andy

Hi Andy,

the non-templated align tries to maximize the overlap between every molecule pairs in the same time. This results in n*(n-1)/2 overlay interaction to maximize. So this scales rather badly and can be used for small molecules if n > 4.Till this time there was no need to think how to speed up alignment if n>2. 

Regarding 3D qsar: There are weight factors for defining how much an atomtype is attracted to an other atomtype during the alignment. like sp3 oxigen to sp3 nitrogen, or aromatic carbon to aromatic nitrogen. In the gui the default is used: 0 for dissimilar and 1 for similar types. The alignment can be trained using these weight factors and hence might directly be used for 3D QSAR in the future...

Cheers

Adrian.