Picture with superposition of molecules

Hi,

I implemented a protocol in which the input is small molecule (like caffeine) that suffers some small flexible modifications. At the end I would like to get a picture, in the HTMl report, where one can see the superposition of both molecules in 3D from some eye perspective. Is it possible in Pipeline Pilot?

Thanks in advance

WS
Answers 6
HP
HP 2009-09-17
Hi WS,

The freely available DS ActiveX control can be used to visualize and rotate 3D molecules within an HTML report. You can download this from our website (it is separate from the Discovery Studio Visualizer, but following the link below and registering will provide access to both)
http://accelrys.com/products/d...-registration-25.php

The attached protocol is a simple example that shows how output molecules can be piped into a reporting component that creates a 'live' 3D window in your HTML report.

You will also want to take a look at the SciTegic Example Protocol:
SciTegic Examples -> Reporting -> Forms and Selectors -> DS Visualizer Display Styles

Please note that since this is ActiveX, it will only work with Windows Internet Explorer.

Tedman
comment imagecomment image
JM
JM 2009-09-23
Werner,

I too, am also very much interested in automated i.e. PP workflow-based, not GUI-based, molecular superposition in both 2D and 3D. Although one might attempt this using alignments using a MCSS, often this is not completely satisfactory.

There can certainly be cases where groups or side-chains can be rotated or adjusted in both 2D and 3D to give improved superposition, where the overlays of the original structures may not be optimal.

If anyone is interested in developing a PP protocol which gives the maximal possible overlay of 2 or more structures in 2D and 3D, especially for molecules with floppy side-chains that would be much appreciated!

Regards,
Jim Metz

Abbott Laboratories
AA
AA 2009-09-24
Certainly 3D alignment is a difficult process with many possible methods and answers. However, one of the strengths of DS is that we have a number of 3D alignment methods available, both in the Visualizer and the DS component collection. For example:

1) Shape based alignment (create a shape from a target conformation and use Ligand Pharmacophore Mapping to align candidate ligands to the shape)
2) Pharmacophore based alignment (Build a Common Features Pharmacophore, the alignments are part of the output)
3) Substructure based alignment (generate conformations then align to a target)
4) Tether based alignment (Define tethers and align)
5) Molecular Field alignment (align based on maximizing electrostatic and steric interactions)

1-3 of these are already available as protocols in one form or another. The last two are available in the client or, by using DiscoveryScript, on the server. Attached is a simple example protocol that flexibly overlays a set of structures to a target. For more information and references see the DS documentation.
comment imagecomment image
HP
HP 2009-09-24
I suspect this is going to be a significantly non-trivial process.

Certainly an optimal way of doing this would have be very useful for 3D QSAR methods (e.g. CoMFA). Given that there was never a great (even good?) method for automating these alignments, I'd be surprised if there was a simple solution.
AA
AA 2009-09-25
Well, I've not really tried. I would think if you have a core it would make things easier, not harder, as there is a pretty good starting point. Is there a public data set we can try this on? I have an idea about using pharmacophores with a core substructure as a required feature I'd like to look into.
HP
HP 2009-09-25
Good point Al. How do these perform in terms of aligning the side chains for a fixed (or approximately fixed) core?

I've got protocols that will apply 4 & 5 by the way, in a slightly different context, so if anyone is interested let me know and I'll extract the relevant bits.