Improving Molecular Pretraining with Complementary Featurizations

TL;DR

This paper introduces MOCO, a molecular pretraining framework that combines multiple complementary featurizations like SMILES, 2D graphs, and 3D geometries, leading to improved performance in molecular property prediction tasks.

Contribution

The paper presents a novel framework, MOCO, that effectively integrates diverse molecular featurizations to enhance pretraining and downstream task performance.

Findings

MOCO outperforms models using single featurizations.

Different featurizations encode distinct chemical information.

Combining featurizations improves molecular property prediction accuracy.

Abstract

Molecular pretraining, which learns molecular representations over massive unlabeled data, has become a prominent paradigm to solve a variety of tasks in computational chemistry and drug discovery. Recently, prosperous progress has been made in molecular pretraining with different molecular featurizations, including 1D SMILES strings, 2D graphs, and 3D geometries. However, the role of molecular featurizations with their corresponding neural architectures in molecular pretraining remains largely unexamined. In this paper, through two case studies -- chirality classification and aromatic ring counting -- we first demonstrate that different featurization techniques convey chemical information differently. In light of this observation, we propose a simple and effective MOlecular pretraining framework with COmplementary featurizations (MOCO). MOCO comprehensively leverages multiple featurizations that complement each other and outperforms existing state-of-the-art models that solely relies on one or two featurizations on a wide range of molecular property prediction tasks.