Modifications in recetox-xMSannotator

Simple Annotation

The input database is expected to contain only the compounds without adduct information (example) and not also the adduct related information (example).
The code computing the matching was rewritten in C++ for performance reasons and also performs the computation of expected mz value for a compound-adduct-pair (see here)

The RT-based clustering as the second step of feature grouping has been reworked as described here
The clustering by this new method results in a finer grouping of peaks into more individual clusters, see the example image below.

The peaks corresponding to the isotopes of a given compound are detected based on several criteria:
1. mass-to-charge ratio
2. mass defect,
3. peak intensity,
4. retention time.
For each annotated feature a peak table is filtered to find peaks that:
1. lie in the same RT cluster as the feature,
2. have RT within rt_tolerance from the feature's RT,
3. have mass defect within mass_defect_tolerance from that of the feature,
After such peaks are identified, their intensities are compared to theoretical expected intensities of corresponding isotopes. Peaks that differ in intensities more than intensity_deviation_tolerance (relative scale) from expected values are filtered out.
Lastly, the detected isotopic peaks are appended to the annotation table