Efficiently predicting high resolution mass spectra with graph neural networks

TL;DR

This paper introduces GrAFF-MS, a graph neural network approach that predicts high-resolution mass spectra efficiently by mapping molecular graphs to formula distributions, significantly improving accuracy and speed in computational metabolomics.

Contribution

The paper presents a novel graph neural network architecture that predicts mass spectra as probability distributions over formulas, overcoming previous tradeoffs between resolution and tractability.

Findings

Achieves lower prediction error than existing methods.

Runs orders of magnitude faster than state-of-the-art approaches.

Uses a fixed vocabulary covering only 2% of observed formulas.

Abstract

Identifying a small molecule from its mass spectrum is the primary open problem in computational metabolomics. This is typically cast as information retrieval: an unknown spectrum is matched against spectra predicted computationally from a large database of chemical structures. However, current approaches to spectrum prediction model the output space in ways that force a tradeoff between capturing high resolution mass information and tractable learning. We resolve this tradeoff by casting spectrum prediction as a mapping from an input molecular graph to a probability distribution over molecular formulas. We discover that a large corpus of mass spectra can be closely approximated using a fixed vocabulary constituting only 2% of all observed formulas. This enables efficient spectrum prediction using an architecture similar to graph classification - GrAFF-MS - achieving significantly lower prediction error and orders-of-magnitude faster runtime than state-of-the-art methods.