The Superhirn FeatureFinderAlgorithm. More...

#include <OpenMS/TRANSFORMATIONS/FEATUREFINDER/FeatureFinderAlgorithmSH.h>

Inheritance diagram for FeatureFinderAlgorithmSH< PeakType, FeatureType >:

Public Types
typedef FeatureFinderAlgorithm < PeakType, FeatureType > ::MapType	MapType
typedef FeatureFinderAlgorithm < PeakType, FeatureType > ::FeatureMapType	FeatureMapType
typedef MapType::SpectrumType	SpectrumType
Public Member Functions
	FeatureFinderAlgorithmSH ()
unsigned int	getNativeScanId (String native_id)
virtual void	run ()
Static Public Member Functions
static FeatureFinderAlgorithm < Peak1D, Feature > *	create ()
static const String	getProductName ()
Protected Attributes
MapType	map_

Detailed Description

template<class PeakType, class FeatureType>
class OpenMS::FeatureFinderAlgorithmSH< PeakType, FeatureType >

The Superhirn FeatureFinderAlgorithm.

The SuperHirn FeatureFinder algorithm is applied by calling "run" on this class, which in turn calls the FeatureFinderAlgorithmSHCtrl to execute the following algorithm:

START by feeding the datavector into startScanParsing (FTPeakDetectController.C) For each scan 1. Centroid it (new CentroidData instance), centroiding is done in CentroidData::calcCentroids 2. Call add_scan_raw_data of ProcessData -> this also does the de-isotoping / feature finding in ProcessData::add_scan_raw_data 3. Apply process_MS1_level_data_structure to the whole map 4. Apply feature merging (MS1FeatureMerger)**, optionally 5. Add to all LC MS/MS runs

Step 2 in ProcessData::add_scan_raw_data works on centroided peaks of a single spectrum 2.1 add to the background intensity controller BackgroundControl::addPeakMSScan which calculates intensity bins 2.2 call "go" on Deisotoper (on single spectrum level) to "de-istope" spectra * 2.3 Converts them to objects of MSPeak type (single spectrum features)

Step 3 works on an instance of ProcessData (clustering de-isotoped peaks from single spectra over RT) and applies the following steps: 3.1 Extract elution peaks (call extract_elution_peaks of ProcessData) 3.2 For all features, it creates a SuperHirn Feature (SHFeature) 3.3 For all features, it computes the elution profile (FeatureLCProfile instance) and adds individual peaks to it

Step 3.1 calls processIntensityMaps from BackgroundController

Deisotoper (Step 2.2) The Deisotoper works on single "peak groups" which is a set of peaks that has a maximal spacing of 1 + exps Da. These peak groups are produced by the CentroidData object [CentroidData::getNextPeakGroup] which internally holds a pointer to the current peak. It basically starts with the first peaks and adds peaks until the next peak is further away than 1+eps The Deisotoper then goes through the peak list, for each charge checks which peaks matches the current charge using IsotopicDist::getMatchingPeaks, creates an instance of DeconvPeak using this mono isotopic charge and then subtracts the current monoisotopic peak from the set of peaks using IsotopicDist::subtractMatchingPeaks (probably to account for overlapping isotopic patterns). Feature Merging Step 4 in MS1FeatureMerger::startFeatureMerging() which calls createMZFeatureClusters(). This method tries to check whether a feature is inside another feature using MS1FeatureMerger::compareMZFeatureBeloning which checks whether the ppm tolerance is below a certain level, the charge state is equal and whether both features have elution profiles.