A Review on Fragment-based De Novo 2D Molecule Generation

TL;DR

This paper reviews fragment-based deep generative models for 2D molecule design, comparing their architectures, performance, limitations, and future research directions in computational chemistry.

Contribution

It provides a comprehensive comparison of existing fragment-based molecular generation models and discusses future research opportunities in the field.

Findings

State-of-the-art models achieve high benchmark performance.

Different fragmentation and generative strategies impact output quality.

Identifies current limitations and promising future research directions.

Abstract

In the field of computational molecule generation, an essential task in the discovery of new chemical compounds, fragment-based deep generative models are a leading approach, consistently achieving state-of-the-art results in molecular design benchmarks as of 2023. We present a detailed comparative assessment of their architectures, highlighting their unique approaches to molecular fragmentation and generative modeling. This review also includes comparisons of output quality, generation speed, and the current limitations of specific models. We also highlight promising avenues for future research that could bridge fragment-based models to real-world applications.