Published December 14, 2019 | Version v2
Poster Open

A learned embedding for efficient joint analysis of millions of mass spectra

  • 1. University of California San Diego, La Jolla, CA, USA
  • 2. University of Washington, Seattle, WA, USA


We propose to train a Siamese neural network using peptide–spectrum assignments to embed spectra in a low-dimensional space such that spectra generated by the same peptide are close to one another. We demonstrate that this learned embedding captures latent properties of the mass spectra, clusters related spectra in the low-dimensional space, and identifies the "dark matter" of the human proteome.



Files (1.3 MB)

Name Size Download all
1.3 MB Preview Download