Improved mass-spectra-based molecule identification using graph neural networks
Detecting and quantifying products of cellular metabolism using Mass Spectrometry (MS) has already shown great promise in many biological and biomedical applications. The biggest challenge in metabolomics is annotation, where measured spectra are assigned chemical identities. Despite advances, current methods provide limited annotation for measured spectra. Here, we explore using graph neural networks (GNNs) to predict the spectra. Our model takes inputs as molecular graphs and predicts spectra intensity at each location. Once the model is trained, for each query of spectra, we use the model to generate spectra predictions for all possible candidate molecules supplied by users and calculate the ranks based on the distance between the predictions and the ground truth. We compare our results to a model that utilizes molecular fingerprints as inputs. Our results show that GNN- based models offer higher performance than fingerprint-based one and these models can be effectively used in untargetted metabolite annotation. Importantly, we show that ranking results heavily depend on the candidate set size and on the similarity of the candidates to the target molecule, thus highlighting the need for consistent, well- characterized evaluation protocols for this domain.
Join the meeting in Sococo VH 102 or Zoom.
Join Zoom Meeting: https://tufts.zoom.us/j/98610939077
PASSWORD: See colloquium email
Dial by your location: +1 646 558 8656 US (New York)
Meeting ID: 986 1093 9077
PASSCODE: See colloquium email