An algorithm to decharge features (i.e. as found by FeatureFinder). More...

#include <OpenMS/ANALYSIS/DECHARGING/FeatureDeconvolution.h>

Inheritance diagram for FeatureDeconvolution:

Public Types
enum	CHARGEMODE { QFROMFEATURE = 1, QHEURISTIC, QALL }
typedef FeatureMap	FeatureMapType
typedef Feature	FeatureType
typedef DPosition< 2 >	ClusterPointType
typedef FeatureMapType::FeatureType::CoordinateType	CoordinateType
typedef ILPDCWrapper::PairsType	PairsType
Public Member Functions
void	compute (const FeatureMapType &fm_in, FeatureMapType &fm_out, ConsensusMap &cons_map, ConsensusMap &cons_map_p)
	Compute a zero-charge feature map from a set of charged features.
Constructors and Destructor s

	FeatureDeconvolution ()
	default constructor
	FeatureDeconvolution (const FeatureDeconvolution &source)
	Copy constructor.
FeatureDeconvolution &	operator= (const FeatureDeconvolution &source)
	Assignment operator.
virtual	~FeatureDeconvolution ()
	destructor
Protected Member Functions
void	updateMembers_ ()
	This method is used to update extra member variables at the end of the setParameters() method.
void	checkSolution_ (const ConsensusMap &cons_map) const
void	inferMoreEdges_ (PairsType &edges, Map< Size, std::set< CmpInfo_ > > &feature_adducts)
void	printEdgesOfConnectedFeatures_ (Size idx_1, Size idx_2, const PairsType &feature_relation)
	A function mostly for debugging.
bool	intensityFilterPassed_ (const Int q1, const Int q2, const Compomer &cmp, const FeatureType &f1, const FeatureType &f2)
	returns true if the intensity filter was passed or switched off
bool	chargeTestworthy_ (const Int feature_charge, const Int putative_charge, const bool other_unchanged) const
	determines if we should test a putative feature charge
Protected Attributes
MassExplainer::AdductsType	potential_adducts_
	List of adducts used to explain mass differences.
Map< Size, String >	map_label_
	labeling table
Map< String, Size >	map_label_inverse_
	labeling table inverse
bool	enable_intensity_filter_
	status of intensity filter for edges
CHARGEMODE	q_try_
	status of charge discovery
Int	verbose_level_
	amount of debug information displayed

Detailed Description

An algorithm to decharge features (i.e. as found by FeatureFinder).

Parameters of this class are:

Name	Type	Default	Restrictions	Description
charge_min	int	1		Minimal possible charge
charge_max	int	10		Maximal possible charge
charge_span_max	int	4	min: 1	Maximal range of charges for a single analyte, i.e. observing q1=[5,6,7] implies span=3. Setting this to 1 will only find adduct variants of the same charge
q_try	string	feature	feature, heuristic, all	Try different values of charge for each feature according to the above settings ('heuristic' [does not test all charges, just the likely ones] or 'all' ), or leave feature charge untouched ('feature').
retention_max_diff	float	1		Maximum allowed RT difference between any two features if their relation shall be determined
retention_max_diff_local	float	1		Maximum allowed RT difference between between two co-features, after adduct shifts have been accounted for (if you do not have any adduct shifts, this value should be equal to 'retention_max_diff', otherwise it should be smaller!)
mass_max_diff	float	0.5		Maximum allowed mass difference [in Th] for a single feature.
potential_adducts	string list	[H+:0.9]		Adducts used to explain mass differences in format: 'Element(+)*:Probability[:RTShift[:Label]]', i.e. the number of '+' indicate the charge, e.g. 'Ca++:0.5' indicates +2. Probabilites have to be in (0,1]. RTShift param is optional and indicates the expected RT shift caused by this adduct, e.g. '(2)H4H-4:1:-3' indicates a 4 deuterium label, which causes early elution by 3 seconds. As a fourth parameter you can add a label which is tagged on every feature which has this adduct. This also determines the map number in the consensus file.
max_neutrals	int	0		Maximal number of neutral adducts(q=0) allowed. Add them in the 'potential_adducts' section!
max_minority_bound	int	2	min: 0	Maximum count of the least probable adduct (according to 'potential_adducts' param) within a charge variant. E.g. setting this to 2 will not allow an adduct composition of '1(H+),3(Na+)' if Na+ is the least probable adduct
min_rt_overlap	float	0.66	min: 0 max: 1	Minimum overlap of the convex hull' RT intersection measured against the union from two features (if CHs are given)
intensity_filter	string	false	true, false	Enable the intensity filter, which will only allow edges between two equally charged features if the intensity of the feature with less likely adducts is smaller than that of the other feature. It is not used for features of different charge.
default_map_label	string	decharged features		Label of map in output consensus file where all features are put by default
verbose_level	int	0	min: 0 max: 3	Amount of debug information given during processing.

Note:

If a section name is documented, the documentation is displayed as tooltip.
Advanced parameter names are italic.

Member Typedef Documentation

typedef DPosition<2> ClusterPointType

typedef FeatureMapType::FeatureType::CoordinateType CoordinateType

typedef FeatureMap FeatureMapType

typedef Feature FeatureType

typedef ILPDCWrapper::PairsType PairsType

Member Enumeration Documentation

enum CHARGEMODE

Enumerator:

QFROMFEATURE
QHEURISTIC
QALL

Constructor & Destructor Documentation

FeatureDeconvolution ( )

default constructor

FeatureDeconvolution ( const FeatureDeconvolution & source )

Copy constructor.

virtual ~FeatureDeconvolution ( ) [virtual]

destructor

Member Function Documentation

bool chargeTestworthy_	(	const Int	feature_charge,
		const Int	putative_charge,
		const bool	other_unchanged
	)			const `[protected]`

determines if we should test a putative feature charge

Answer query given the internal status of q_try. Features with q<=0 always return true.

void checkSolution_ ( const ConsensusMap & cons_map ) const [protected]

void compute	(	const FeatureMapType &	fm_in,
		FeatureMapType &	fm_out,
		ConsensusMap &	cons_map,
		ConsensusMap &	cons_map_p
	)

Compute a zero-charge feature map from a set of charged features.

Find putative ChargePairs, then score them and hand over to ILP.

Parameters:

	fm_in	Input feature-map
	fm_out	Output feature-map (sorted by position and augmented with user params)
	cons_map	[out] Output of grouped features belonging to a charge group
	cons_map_p	[out] Output of paired features connected by an edge

void inferMoreEdges_	(	PairsType &	edges,
		Map< Size, std::set< CmpInfo_ > > &	feature_adducts
	)			`[protected]`

test if "simple" edges have alternative (more difficult explanation) supported by neighboring edges e.g. (.) -> (H+) might be augmented to (Na+) -> (H+Na+)