Tests a group of data points in an LC-MS map for goodness-of-fit with a 2D averagine model. More...

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

Inheritance diagram for ModelFitter< PeakType, FeatureType >:

Public Types
enum	{ RT = Peak2D::RT, MZ = Peak2D::MZ }
typedef IndexSet::const_iterator	IndexSetIter
	IndexSet iterator.
typedef Feature::QualityType	QualityType
	Quality of a feature.
typedef Feature::CoordinateType	CoordinateType
	Single coordinate.
typedef Feature::IntensityType	IntensityType
	Single intensity.
typedef Feature::ChargeType	ChargeType
	Isotope charge.
typedef FeaFiModule< PeakType, FeatureType >	Base
	FeaFiModule.
typedef std::vector< PeakType >	RawDataArrayType
	Raw data container type using for the temporary storage of the input data.
Public Member Functions
	ModelFitter (const MSExperiment< PeakType > map, FeatureMap< FeatureType > features, FeatureFinder *ff)
	Constructor.
virtual	~ModelFitter ()
	Destructor.
void	setMonoIsotopicMass (CoordinateType mz)
	Sets or fixed the monoisotopic m/z at a specific position.
Feature	fit (const ChargedIndexSet &index_set)
	Return next feature.
Protected Member Functions
virtual void	updateMembers_ ()
QualityType	fitMZLoop_ (const ChargedIndexSet &set, const ChargeType &charge)
	main fit loop
QualityType	evaluate_ (const IndexSet &set) const
	evaluate 2d-model
QualityType	fitRT_ (InterpolationModel *&model) const
	1d fit in RT
QualityType	fitMZ_ (InterpolationModel *&model, const ChargeType &charge) const
	1d fit in MZ
void	doProjectionDim_ (const Int dim, const ChargedIndexSet &index_set, RawDataArrayType &set) const
	Project the raw data into 1-dim array.
Protected Attributes
ProductModel< 2 >	model2D_
	2D model
Math::BasicStatistics	mz_stat_
	statistics for mz
Math::AsymmetricStatistics	rt_stat_
	statistics for rt
RawDataArrayType	mz_input_data_
	mz raw data
RawDataArrayType	rt_input_data_
	rt raw data
CoordinateType	tolerance_stdev_box_
	tolerance used for bounding box
CoordinateType	monoisotopic_mz_
	monoistopic mass
CoordinateType	interpolation_step_mz_
	interpolation step size (in m/z)
CoordinateType	interpolation_step_rt_
	interpolation step size (in retention time)
Int	max_isotope_
	maximum isotopic rank to be considered
CoordinateType	iso_stdev_first_
	first stdev
CoordinateType	iso_stdev_last_
	last stdev
CoordinateType	iso_stdev_stepsize_
	step size
Int	first_mz_model_
	first mz model (0=Gaussian, 1....n = charge )
Int	last_mz_model_
	last mz model
CoordinateType	isotope_stdev_
	isotope stdev
String	algorithm_
	algorithm
Int	max_iteration_
	Maximum number of iterations.
CoordinateType	deltaAbsError_
	Absolute error.
CoordinateType	deltaRelError_
	Relative error.
Math::BasicStatistics	basic_stat_
	statistics
QualityType	quality_rt_
	fit quality in RT dimension
QualityType	quality_mz_
	fit quality in MZ dimension
Private Member Functions
	ModelFitter ()
	Not implemented.
ModelFitter &	operator= (const ModelFitter &)
	Not implemented.
	ModelFitter (const ModelFitter &)
	Not implemented.

Detailed Description

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

Tests a group of data points in an LC-MS map for goodness-of-fit with a 2D averagine model.

The ModelFitter class uses a gaussian or isotope model in m/z and bigauss, lmagauss (bigauss with Levenberg-Marquardt) or emg (exponentially modified Gaussian with lma optimized parameters) in retention time.

For the isotope model, we test different charge states and peaks widths.

Todo:: Fix problem with other peak types than Peak1D and enable the respective tests for FeatureFinderAlgorithmSimple, FeatureFinderAlgorithmSimplest, FeatureFinderAlgorithmWavelet (Clemens)

Parameters of this class are:

Name	Type	Default	Restrictions	Description
fit_algorithm	string	simple	simple, simplest, wavelet	Fitting algorithm type (internal parameter).
max_iteration	int	500	min: 1	Maximum number of iterations for fitting with Levenberg-Marquardt algorithm.
deltaAbsError	float	0.0001	min: 0	Absolute error used by the Levenberg-Marquardt algorithm.
deltaRelError	float	0.0001	min: 0	Relative error used by the Levenberg-Marquardt algorithm.
tolerance_stdev_bounding_box	float	3	min: 0	Bounding box has range [minimim of data, maximum of data] enlarged by tolerance_stdev_bounding_box times the standard deviation of the data
intensity_cutoff_factor	float	0.0500000007450581	min: 0 max: 1	Cutoff peaks with a predicted intensity below intensity_cutoff_factor times the maximal intensity of the model
feature_intensity_sum	int	1	min: 0 max: 1	Determines what is reported as feature intensity. 1: the sum of peak intensities; 0: the maximum intensity of all peaks
min_num_peaks:final	int	5	min: 1	Minimum number of peaks left after cutoff. If smaller, feature will be discarded.
min_num_peaks:extended	int	10	min: 1	Minimum number of peaks after extension. If smaller, feature will be discarded.
rt:interpolation_step	float	0.200000002980232	min: 0	Step size in seconds used to interpolate model for RT.
mz:interpolation_step	float	0.0299999993294477	min: 0.001	Interpolation step size for m/z.
mz:model_type:first	int	1	min: 0	Numeric id of first m/z model fitted (usually indicating the charge state), 0 = no isotope pattern (fit a single gaussian).
mz:model_type:last	int	4	min: 0	Numeric id of last m/z model fitted (usually indicating the charge state), 0 = no isotope pattern (fit a single gaussian).
quality:type	string	Correlation	Correlation, RankCorrelation	Type of the quality measure used to assess the fit of model vs data.
quality:minimum	float	0.649999976158142	min: 0 max: 1	Minimum quality of fit, features below this threshold are discarded.
isotope_model:stdev:first	float	0.0399999991059303	min: 0	First standard deviation to be considered for isotope model.
isotope_model:stdev:last	float	0.119999997317791	min: 0	Last standard deviation to be considered for isotope model.
isotope_model:stdev:step	float	0.0399999991059303	min: 0	Step size for standard deviations considered for isotope model.
isotope_model:averagines:C	float	0.0444398894906044	min: 0	Number of C atoms per Dalton of the mass.
isotope_model:averagines:H	float	0.0698157176375389	min: 0	Number of H atoms per Dalton of the mass.
isotope_model:averagines:N	float	0.0122177302837372	min: 0	Number of N atoms per Dalton of the mass.
isotope_model:averagines:O	float	0.0132939899340272	min: 0	Number of O atoms per Dalton of the mass.
isotope_model:averagines:S	float	0.000375250005163252	min: 0	Number of S atoms per Dalton of the mass.
isotope_model:isotope:trim_right_cutoff	float	0.00100000004749745	min: 0	Cutoff for averagine distribution, trailing isotopes below this relative intensity are not considered.
isotope_model:isotope:maximum	int	100	min: 1	Maximum number of isotopes being used for the IsotopeModel.
isotope_model:isotope:distance	float	1.00049495697021	min: 0	Distance between consecutive isotopic peaks.

Note:

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

Member Typedef Documentation

typedef FeaFiModule<PeakType, FeatureType> Base

FeaFiModule.

typedef Feature::ChargeType ChargeType

Isotope charge.

typedef Feature::CoordinateType CoordinateType

Single coordinate.

typedef IndexSet::const_iterator IndexSetIter

IndexSet iterator.

typedef Feature::IntensityType IntensityType

Single intensity.

typedef Feature::QualityType QualityType

Quality of a feature.

typedef std::vector<PeakType> RawDataArrayType

Raw data container type using for the temporary storage of the input data.

Member Enumeration Documentation

anonymous enum

Enumerator:

RT
MZ

Constructor & Destructor Documentation

ModelFitter	(	const MSExperiment< PeakType > *	map,
		FeatureMap< FeatureType > *	features,
		FeatureFinder *	ff
	)			`[inline]`

Constructor.

References StringList::create(), DefaultParamHandler::defaults_, DefaultParamHandler::defaultsToParam_(), Param::setMaxFloat(), Param::setMaxInt(), Param::setMinFloat(), Param::setMinInt(), DefaultParamHandler::setName(), Param::setSectionDescription(), Param::setValidStrings(), and Param::setValue().

virtual ~ModelFitter ( ) [inline, virtual]

Destructor.

ModelFitter ( ) [private]

Not implemented.

ModelFitter ( const ModelFitter< PeakType, FeatureType > & ) [private]

Not implemented.

Member Function Documentation

void doProjectionDim_	(	const Int	dim,
		const ChargedIndexSet &	index_set,
		RawDataArrayType &	set
	)			const `[inline, protected]`

Project the raw data into 1-dim array.

References ModelFitter< PeakType, FeatureType >::algorithm_, FeaFiModule< PeakType, FeatureType >::getPeakIntensity(), FeaFiModule< PeakType, FeatureType >::getPeakMz(), FeaFiModule< PeakType, FeatureType >::getPeakRt(), and ModelFitter< PeakType, FeatureType >::MZ.

Referenced by ModelFitter< PeakType, FeatureType >::fit().

QualityType evaluate_ ( const IndexSet & set ) const [inline, protected]

evaluate 2d-model

References ModelFitter< PeakType, FeatureType >::algorithm_, ProductModel< 2 >::getIntensity(), FeaFiModule< PeakType, FeatureType >::getPeakIntensity(), FeaFiModule< PeakType, FeatureType >::getPeakMz(), Param::getValue(), ModelFitter< PeakType, FeatureType >::model2D_, DefaultParamHandler::param_, OpenMS::Math::pearsonCorrelationCoefficient(), and OpenMS::Math::rankCorrelationCoefficient().

Referenced by ModelFitter< PeakType, FeatureType >::fitMZLoop_().

Feature fit ( const ChargedIndexSet & index_set ) [inline]

Return next feature.

Exceptions:

	Exception::UnableToFit	is thrown if fitting cannot be performed
	Exception::InvalidParameter	if first and last charge to test do not define a range (first<=last)

References FeaFiModule< PeakType, FeatureType >::addConvexHull(), IsotopeCluster::ChargedIndexSet::charge, ModelDescription< D >::createModel(), ModelFitter< PeakType, FeatureType >::doProjectionDim_(), Feature::encloses(), FeaFiModule< PeakType, FeatureType >::ff_, ModelFitter< PeakType, FeatureType >::first_mz_model_, ModelFitter< PeakType, FeatureType >::fitMZLoop_(), ModelFitter< PeakType, FeatureType >::fitRT_(), ConvexHull2D::getBoundingBox(), BaseFeature::getCharge(), Feature::getConvexHull(), ProductModel< 2 >::getIntensity(), ProductModel< 2 >::getModel(), Feature::getModelDescription(), Peak2D::getMZ(), DefaultParamHandler::getName(), Feature::getOverallQuality(), FeatureFinder::getPeakFlag(), FeaFiModule< PeakType, FeatureType >::getPeakIntensity(), FeaFiModule< PeakType, FeatureType >::getPeakMz(), FeaFiModule< PeakType, FeatureType >::getPeakRt(), Feature::getQuality(), Peak2D::getRT(), Feature::getSubordinates(), Param::getValue(), BaseModel< D >::isContained(), ModelFitter< PeakType, FeatureType >::last_mz_model_, ModelFitter< PeakType, FeatureType >::model2D_, ModelFitter< PeakType, FeatureType >::MZ, ModelFitter< PeakType, FeatureType >::mz_input_data_, ModelFitter< PeakType, FeatureType >::mz_stat_, DefaultParamHandler::param_, ModelFitter< PeakType, FeatureType >::quality_mz_, ModelFitter< PeakType, FeatureType >::quality_rt_, ModelFitter< PeakType, FeatureType >::RT, ModelFitter< PeakType, FeatureType >::rt_input_data_, ModelFitter< PeakType, FeatureType >::rt_stat_, BaseFeature::setCharge(), BaseModel< D >::setCutOff(), Peak2D::setIntensity(), MetaInfoInterface::setMetaValue(), ProductModel< 2 >::setModel(), Feature::setModelDescription(), Peak2D::setMZ(), Feature::setOverallQuality(), Feature::setQuality(), Peak2D::setRT(), ProductModel< 2 >::setScale(), AsymmetricStatistics< Real >::update(), and BasicStatistics< RealT >::update().

Referenced by FeatureFinderAlgorithmSimplest< PeakType, FeatureType >::run(), and FeatureFinderAlgorithmSimple< PeakType, FeatureType >::run().

QualityType fitMZ_	(	InterpolationModel *&	model,
		const ChargeType &	charge
	)			const `[inline, protected]`

1d fit in MZ

Referenced by ModelFitter< PeakType, FeatureType >::fitMZLoop_().

QualityType fitMZLoop_	(	const ChargedIndexSet &	set,
		const ChargeType &	charge
	)			`[inline, protected]`

main fit loop

Referenced by ModelFitter< PeakType, FeatureType >::fit().

QualityType fitRT_ ( InterpolationModel *& model ) const [inline, protected]

1d fit in RT

Referenced by ModelFitter< PeakType, FeatureType >::fit().

ModelFitter& operator= ( const ModelFitter< PeakType, FeatureType > & ) [private]

Not implemented.

void setMonoIsotopicMass ( CoordinateType mz ) [inline]

Sets or fixed the monoisotopic m/z at a specific position.

Parameters:

mz

The monoisotopic m/z that occurres in the current data set.

References ModelFitter< PeakType, FeatureType >::monoisotopic_mz_.

virtual void updateMembers_ ( ) [inline, protected, virtual]

Member Data Documentation

String algorithm_ [protected]

algorithm

Referenced by ModelFitter< PeakType, FeatureType >::doProjectionDim_(), ModelFitter< PeakType, FeatureType >::evaluate_(), ModelFitter< PeakType, FeatureType >::fitMZ_(), ModelFitter< PeakType, FeatureType >::fitRT_(), and ModelFitter< PeakType, FeatureType >::updateMembers_().

Math::BasicStatistics basic_stat_ [protected]

statistics

CoordinateType deltaAbsError_ [protected]

Absolute error.

Test for the convergence of the sequence by comparing the last iteration step dx with the absolute error epsabs and relative error epsrel to the current position x

Referenced by ModelFitter< PeakType, FeatureType >::fitMZ_(), ModelFitter< PeakType, FeatureType >::fitRT_(), and ModelFitter< PeakType, FeatureType >::updateMembers_().

CoordinateType deltaRelError_ [protected]

Relative error.

Referenced by ModelFitter< PeakType, FeatureType >::fitMZ_(), ModelFitter< PeakType, FeatureType >::fitRT_(), and ModelFitter< PeakType, FeatureType >::updateMembers_().

Int first_mz_model_ [protected]

first mz model (0=Gaussian, 1....n = charge )

Referenced by ModelFitter< PeakType, FeatureType >::fit(), and ModelFitter< PeakType, FeatureType >::updateMembers_().

CoordinateType interpolation_step_mz_ [protected]

interpolation step size (in m/z)

Referenced by ModelFitter< PeakType, FeatureType >::fitMZ_(), and ModelFitter< PeakType, FeatureType >::updateMembers_().

CoordinateType interpolation_step_rt_ [protected]

interpolation step size (in retention time)