Meaning of RDKit descriptors

[anne]

Hello,

I am not a chemist and would like a brief description of what the RDKit descriptors mean. For example, what is peoe_VSA9? Layman's terms would be nice. Where can I find that?

thanks!

[Aaron_Hart]

I am not sure, but it looks like they are based descriptors defined in a book titled:

"Derivation and Applications of Molecular Descriptors Based on Approximate Surface Area"

It is available from Springer (Along with ISBN and DOI data):

http://link.springer.com/protocol/10.1385%2F1-59259-802-1%3A261

[greglandrum]

Hi Anne,

There are descriptions of many of the RDKit descriptors, along with links to the literature for some, here: http://www.rdkit.org/docs/GettingStartedInPython.html#list-of-available-descriptors

Looking at that list, I see that it's been a while since it was updated. 

To answer your specific question: the  _VSA descriptors are members of a descriptor class that can be quite useful for building statistical models but that are essentially uninterpretable.[1]

Best,

-greg

[1] This isn't strictly true, if one invests considerable time in it, it's possible to come up with extremely hand-wavy interpretations. In my experience the effort has almost never been worth it.