DeepMass is a suite of tools in development at Verily to enable mass spectrometry data analysis using modern machine learning techniques.
The first tool to be included in DeepMass is Prism, a computational framework to predict MS2 fragmentation peak intensities. Developed in collaboration with the Max Planck Institute, Prism's accuracy approaches the theoretical limit imposed by experimental uncertainty.
We provide Prism as a service using Google Cloud Machine Learning Engine. This repository contains instructions and utilities to get started.
We plan on expanding & updating DeepMass into a comprehensive suite of tools for a broad array of Mass Spec capabilities. For information, updates, or to contribute, please reach out to us at DeepMass@google.com.