DOGS: Reaction-Driven de novo Design of Bioactive Compounds

Abstract

We present a computational method for the reaction-based de novo design of drug-like molecules. The software DOGS (Design of Genuine Structures) features a ligand-based strategy for automated ‘in silico’ assembly of potentially novel bioactive compounds. The quality of the designed compounds is assessed by a graph kernel method measuring their similarity to known bioactive reference ligands in terms of structural and pharmacophoric features. We implemented a deterministic compound construction procedure that explicitly considers compound synthesizability, based on a compilation of 25'144 readily available synthetic building blocks and 58 established reaction principles. This enables the software to suggest a synthesis route for each designed compound. Two prospective case studies are presented together with details on the algorithm and its implementation. De novo designed ligand candidates for the human histamine H4 receptor and γ-secretase were synthesized as suggested by the software. The computational approach proved to be suitable for scaffold-hopping from known ligands to novel chemotypes, and for generating bioactive molecules with drug-like properties.

Identifying the macromolecular targets of de novo-designed chemical entities through self-organizing map consensus

“…As part of an early discovery program, we generated innovative molecules using our ligand-based de novo design software DOGS (inSili.com LLC) (19). The process by which DOGS generates a compound uses virtual organic synthesis.…”

IADE: a system for intelligent automatic design of bioisosteric analogs

“…Therefore the processing is very fast. As seen from the previous description, the IADE design strategy is not transformation driven, as for example the program DOGS [22] that uses a database of common chemical reactions to design new molecules, but fragment driven. In the next step, the retained analogs are converted into 3D by CORINA [23].…”

Applicability Domains and Consistent Structure Generation

“…[3,4] Using the QSPR and QSAR models, we can estimate values of objective variables, y, such as activities or properties from values of explanatory variables or molecular descriptors, X, by modeling relationships between X and y using compounds with known activities or properties, then inputting new data for X into the QSPR and QSAR models. New data for X come from virtual molecular structures created by structure generators [5][6][7][8] . Thus, virtual molecular structures are input into the QSAR or QSPR models, and activities or properties are estimated for each.…”

“…Based on this flow of de novo molecular design, we can effectively design drugs and highly functional products. Table 1 shows recent examples of molecules that were designed de novo using the QSAR or QSPR models [6,[9][10][11][12][13][14][15] . One of the important points in de novo molecular design is that estimated activity or property values are reliable only for virtual molecular structures that are similar to those of the training compounds, whereas any virtual molecular structures can be input into the QSAR or QSPR models.…”