A retrosynthetic analysis algorithm implementation

Abstract

The need for synthetic route design arises frequently in discovery-oriented chemistry organizations. While traditionally finding solutions to this problem has been the domain of human experts, several computational approaches, aided by the algorithmic advances and the availability of large reaction collections, have recently been reported. Herein we present our own implementation of a retrosynthetic analysis method and demonstrate its capabilities in an attempt to identify synthetic routes for a collection of approved drugs. Our results indicate that the method, leveraging on reaction transformation rules learned from a large patent reaction dataset, can identify multiple theoretically feasible synthetic routes and, thus, support research chemist everyday efforts.

Keywords: Chemical synthesis; Reaction informatics; Retrosynthetic analysis; Synthetic route design.

Fig. 1

a A reaction example from US Patent 04703036 and b the reversed version of the same reaction. The reaction core is highlighted in red on the product side

Fig. 2

A heterocycle forming reaction from US patent number US20030149264A1. The reaction core consists of all numbered atoms 1, 2, 3, 4, 5

Fig. 3

Reaction clusters defined using smiles-like signatures highlighted in red (radius 0), red and green (radius 1) and red, green and purple (radius 2). Note the effect of the atom properties used to the classification produced

Fig. 4

The RTSA-train process

Fig. 5

Pseudocode describing the high level RTSA-Design process. Note that step 2 is optional and could be omitted. Alternatively, the process may be supplemented with a recursive RTSA search for each of the synthons not found to be available leading to a multi-step synthesis route

Fig. 6

Fraction of reaction examples covered as a function of number of reaction signatures for different radii sizes. Note the steep curves indicating that a large fraction of the reactions examined is covered by relatively few reaction signatures

Fig. 7

Example RRT’s extracted using signatures of radius 0 and support 10,000

Fig. 8

Examples of reverse reaction templates exemplifying the 4 signatures of radius 0 with the highest support in Dataset 1. Note that the templates are reversed, i.e. the original products are to the left-hand side of the figure

Fig. 9

DrugBankID:DB00619 and four potential synthesis routes identified by RTSA using reverse reaction templates of signature radius 2 and frequency 100 extracted from Dataset 1. Note that the origin of the reverse reaction template used to define the route is shown below each pair of synthons