Venn Diagram from Scientific Research

Even though the number of publicly available unannotated experimental mass spectra has been rapidly growing over recent years (left; green), annotated spectral libraries are still highly limited in terms of both the number of spectra (left; pink) and the coverage of molecular structures (right; Venn diagram). State-of-the-art annotation methods rely on spectral libraries as training or retrieval datasets. By contrast, we base our method on training from vast unannotated MS/MS datasets, assuming that the molecular coverage of these data surpasses spectral libraries (right; dashed green shape). K, thousand.

DOI