Mass spectrometry (MS) is gaining momentum in various fields especially forensic applications and fundamental science research. As a result, the need for fast and reliable MS data analyses is growing rapidly. Conventional methods are heavily relying on expert knowledge, and suitable for relatively small data size. An alternative for traditional machine learning and data analyses methods are deep neural networks and one of the widely used types of these networks are convolutional neural networks (CNN).
In this study, we present an innovative MS analyses pipeline which is powered by CNNs for pattern recognition. We show how this pipeline can speed up and improve the accuracy of data analysing.
Mehrdad Jahanbanifard, Naturalis