v.0.2.0-beta

We're happy to announce the first release of peakonly with a graphical user interface.

Major improvements/changes:

• convenient interface for processing and visualizing data, as well as data annotations and more;
• ability to process multiple files (batch), including peaks matching among different files and automated zero fillings;
• increased annotated dataset (now ~6000 ROI), including data from GC-MS;
• the number of classes decreased from three (noise, peaks, uncertain peaks) to two (noise, peaks) in order to simplify a process of data annotation for re-training of CNNs.

Minor improvements/changes:

• the process of ROI detection is now two times faster;
• the possibility to accelerate the process of ROI detection even further (up to ~30x), but a code compilation is required;
• the ability to process MS2 data (only MS1 spectra are taken into account);

The current release includes experimental features (for example, ROI matching among different files and possibility to use recurrent CNN), which may possibly dramatically change in the near future.

Initial release

Initial release matching the published article: Deep Learning for the Precise Peak Detection in High-Resolution LC–MS Data, Arsenty D. Melnikov, Yuri P. Tsentalovich, and Vadim V. Yanshole, Analytical Chemistry 2020 92 (1), 588-592, DOI: 10.1021/acs.analchem.9b04811