Conclusions: The assessment of a panel of equations for the prediction of serum osmolarity led to identification of one formula with a greater diagnostic performance.
However, graphic representations intended for easy identification of the formula and for illustrating methods of preparation will still be permitted.
We evaluate FTs using two derived measures: both for the identification of molecular formulas of unknown compounds, and for searching a database for chemically similar compounds, the new FTs perform significantly better than state-of-the-art methods.
We demonstrate the superiority of the new scoring for two tasks: both for the de novo identification of molecular formulas of unknown compounds, and for searching a database for structurally similar compounds, our method SIRIUS 3, performs significantly better than the previous version of our method, as well as other methods for this task.
Thus, FTICR MS allows for the identification of molecular formulae with increased confidence.
A key deterrent to the identification of chemical formulae has historically stemmed from degeneracy, which increases with mass.
We demonstrate here that the SIL method is better than existing methods at identification of chemical formulae for metabolites larger than 500 Da.
Samples were measured both in linear mode, providing information on the total number of different structures, and in reflector mode, for identification of molecular formulas based on precise mass measurements.
Stable isotope labeling methods allow counting of C and N per formula unit and can lead to identification of the chemical formula without reliance on the natural isotopic abundance patterns and without using a restricted chemical formula generator.
Different from supervised Machine Learning, hyperparameters of SIRIUS 3 are not trained in a way that maximizes, say, correct identifications of molecular formulas.
