GraphRT: A graph-based deep learning model for predicting the retention time of peptides

TL;DR

GraphRT is a novel deep learning model that combines graph neural networks and recurrent neural networks to accurately predict peptide retention times in LC MSMS, including peptides with unseen modifications.

Contribution

It introduces a dual graph and sequence-based approach for peptide RT prediction, outperforming existing models and handling unseen modifications.

Findings

Outperforms all current state-of-the-art models.

Accurately predicts retention times for peptides with unseen modifications.

Integrates atomic-level graph features with sequence context.

Abstract

GraphRT is a graph based deep learning model that predicts the retention time (RT) of peptides in liquid chromatography tandem mass spectrometry (LC MSMS) experiments. Each amino acid is represented as a graph, capturing its atomic and structural properties through a graph neural network. This enables the model to understand not just the chemical composition of each amino acid, but also the intricate relationships between its atoms. The sequential context of the peptide the order and interaction of amino acids in the sequence is then encoded using recurrent neural networks. This dual approach of graph based and sequential modelling allows for a comprehensive understanding of both the individual characteristics of amino acids and their collective behaviour in a peptide sequence. GraphRT outperforms all current state of the art models and can predict retention time for peptides containing unseen modifications.