Fragment assembly in the automated molecular invention system: INVENTON.

摘要

Our central hypothesis is that computers, when properly programmed, can be more creative than chemists at inventing molecular structures. The INVENTON project is being developed to this end. The INVENTON paradigm places the role of creativity in the hands of the computer. The chemist describes the design problem by specifying objectives and constraints. INVENTON analyzes the problem and determines its own course of action. INVENTON carries itself through that course to the invention of a population of molecular structures. The chemist can then analyze the results. INVENTON works unsupervised, and we therefore refer to it as an automated molecular invention system.;The present work describes the development of INVENTON's first method of molecular invention, FASM. FASM is a fragment assembly based method that utilizes a small library of fragments and principles of stable molecular structure to assemble fragments into larger, more complex structures. It works in a strictly goal-directed fashion, assembling fragments in a stepwise manner until all objectives and criteria are met. The methodology is applied to three unrelated problems: morphine mimics, self-assembling nanotubes, and HIV protease inhibitors.;FASM is found to be a very successful molecular invention methodology. A key to the creativity and productivity of the method is due to internal knowledge of stable organic structure, which allows it to resolve temporal problems on the fly that would have otherwise lead to tactical failure. The results of the problems investigated herein demonstrate true creativity in molecular structure. Several structures were submitted for patent. Morphine was reinvented.