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.