A transformer model forde novosequencing of data-independent acquisition mass spectrometry data

Abstract

A core computational challenge in the analysis of mass spectrometry data is thede novosequencing problem, in which the generating amino acid sequence is inferred directly from an observed fragmentation spectrum without the use of a sequence database. Recently, deep learning models have made significant advances inde novosequencing by learning from massive datasets of high-confidence labeled mass spectra. However, these methods are primarily designed for data-dependent acquisition (DDA) experiments. Over the past decade, the field of mass spectrometry has been moving toward using data-independent acquisition (DIA) protocols for the analysis of complex proteomic samples due to their superior specificity and reproducibility. Hence, we present a newde novosequencing model called Cascadia, which uses a transformer architecture to handle the more complex data generated by DIA protocols. In comparisons with existing approaches forde novosequencing of DIA data, Cascadia achieves state-of-the-art performance across a range of instruments and experimental protocols. Additionally, we demonstrate Cascadia’s ability to accurately discoverde novocoding variants and peptides from the variable region of antibodies.