From: root (root_at_liposome.genebee.msu.su)
Date: Tue Dec 02 1997 - 06:18:33 CST

On Tue, 2 Dec 1997, SciFarma wrote:

> Please, give me ideas on the way to represent "graphs" (i.e. chemical
> structures) in a linear form that can be optimized by genetic algorithms .

I believe the SMILES system is already suitable for that. According to
Gasteiger's claims they can generate trivially minimized 3d structures
nearly in realtime. For GA purposes, you might consider including a
mutation frequency map, and a recursive macro extension of the SMILES
system.

http://www.daylight.com/
http://schiele.organik.uni-erlangen.de/corina/free_struct.html

Regards,
Eugene

> **********E-mai to:SALVGUC_at_mail.tau.it************
>