openmolecules.org

 
Home » DataWarrior » Functionality » Display Full Similarity Matrix
Display Full Similarity Matrix [message #140] Thu, 07 January 2016 19:20 Go to next message
caius is currently offline  caius
Messages: 2
Registered: January 2016
Junior Member
Is it possible to display the full similarity matrix, either as a separate document or within the program for export? The manual states that the matrix is calculated completely anyway when conducting similarity analysis, but the "structure pair document" only displays information on compounds with similarity > 0.8
Re: Display Full Similarity Matrix [message #143 is a reply to message #140] Sun, 10 January 2016 23:45 Go to previous messageGo to next message
thomas is currently offline  thomas
Messages: 646
Registered: June 2014
Senior Member
No, currently there is no way to write the entire similarity matrix into a file, although it is created internally for various purposes. It never seemed obvious to me that there would be a need. If you are able to program in Java, you may download the source code and write a small program to calculate the matrix. This would only need a few lines of code. Let me know, if I can be of help with that.

Thomas
Re: Display Full Similarity Matrix [message #144 is a reply to message #143] Mon, 11 January 2016 22:01 Go to previous messageGo to next message
caius is currently offline  caius
Messages: 2
Registered: January 2016
Junior Member
I was able to approximate matrix output using macros and the calculated column generation, but the results from chemsim(Flexophore_of_Structure, idcode) don't match the Structure Similarity values available in the similarity chart or the Structure[Structure List, Similarity] filter. For instance, use of chemsim does not yield a self similarity score of 1. Is there any way to get the outputs of "Structure Similarity" or "Structure (is similar to Flexophore)" into the table?
Re: Display Full Similarity Matrix [message #145 is a reply to message #144] Fri, 15 January 2016 22:45 Go to previous message
thomas is currently offline  thomas
Messages: 646
Registered: June 2014
Senior Member
The Flexophore is the only descriptor that give self-similarity values slightly below 1.0,
because under the hood the Flexophore is generated from a fresh set of representative conformers,
which every time look a little different. Nevertheless, similarity values from two different
Flexophore descriptors should usually be above 0.99 if both were generated from the same molecule.
Structure filters, chemsim() function or graphical view axis all use the same flexophore similarity
calculation, but have the same issue that there is a slight variation in similarity values, if the
flexophore is calculated multiple times.
Previous Topic: How to avoid shuffling of columns after importing?
Next Topic: Drawing Multi-step reaction schemes.
Goto Forum:
  


Current Time: Thu Mar 28 21:49:51 CET 2024

Total time taken to generate the page: 0.08109 seconds