UC Riverside

Mass spectrometry (MS) is a foundational element in contemporary analytical chemistry, facilitating biomolecule identification and playing a pivotal role in deciphering the intricacies of biological systems. Within the MS field, tandem mass spectra (MS/MS) is vital in the process of identifying molecules in untargeted approaches. MS provides structural information about the mass to charge (m/z) of the molecules under analysis. MS/MS provides the fragmentation pattern of the molecules under study, providing more information about the molecule and enhancing the proper identification. Nevertheless, the availability of experimental MS/MS from compounds is limited due to the lack of time, money, and/or availability of reference standards. In those cases, the prediction of mass spectra enables the identification of molecules that have not been experimentally analyzed before, and they are especially important for de-novo identification. However, the current prediction methods still have limitations. QCxMS (Quantum Chemical Mass Spectrometry), the sole tool for predicting mass spectra using dynamic molecular methods, looks like a promising alternative for improving the prediction of fragmentation patterns of molecules. The current experiments using QCxMS have shown limited results. QCxMS has long grappled with issues of time inefficiency and accuracy. This thesis introduces an innovative approach to mitigate these challenges employing a Nextflow workflow to parallelize computations on clusters. The workflow reduces the computation time and facilitates the execution of large-scale experiments and evaluations, thus enabling the possibility of improving the precision and recall predictions of QCxMS by tuning the setup of the simulations. Beyond efficiency enhancements, extensive experiments were conducted to assess the predictive capabilities of QCxMS, identifying the parameters that outperform the default settings of QCxMS.