This class provides the two-dimensional optimization of the picked peak parameters. More...

#include <OpenMS/TRANSFORMATIONS/RAW2PEAK/TwoDOptimization.h>

Inheritance diagram for TwoDOptimization:

Classes
struct	Data
	Helper struct (contains the size of an area and a raw data container). More...
Public Member Functions
	TwoDOptimization ()
	Constructor.
	TwoDOptimization (const TwoDOptimization &opt)
	Copy constructor.
virtual	~TwoDOptimization ()
	Destructor.
TwoDOptimization &	operator= (const TwoDOptimization &opt)
	Assignment operator.
DoubleReal	getMZTolerance () const
	Non-mutable access to the matching epsilon.
void	setMZTolerance (DoubleReal tolerance_mz)
	Mutable access to the matching epsilon.
DoubleReal	getMaxPeakDistance () const
	Non-mutable access to the maximal peak distance in a cluster.
void	setMaxPeakDistance (DoubleReal max_peak_distance)
	Mutable access to the maximal peak distance in a cluster.
DoubleReal	getMaxAbsError () const
	Non-mutable access to the maximal absolute error.
void	setMaxAbsError (DoubleReal eps_abs)
	Mutable access to the maximal absolute error.
DoubleReal	getMaxRelError () const
	Non-mutable access to the maximal relative error.
void	setMaxRelError (DoubleReal eps_rel)
	Mutable access to the maximal relative error.
UInt	getMaxIterations () const
	Non-mutable access to the maximal number of iterations.
void	setMaxIterations (UInt max_iteration)
	Mutable access to the maximal number of iterations.
const OptimizationFunctions::PenaltyFactorsIntensity &	getPenalties () const
	Non-mutable access to the minimal number of adjacent scans.
void	setPenalties (OptimizationFunctions::PenaltyFactorsIntensity &penalties)
	Mutable access to the minimal number of adjacent scans.
template<typename InputSpectrumIterator , typename OutputPeakType >
void	optimize (InputSpectrumIterator first, InputSpectrumIterator last, MSExperiment< OutputPeakType > &ms_exp, bool real2D=true)
	Find two dimensional peak clusters and optimize their peak parameters.
Protected Member Functions
void	updateMembers_ ()
	update members method from DefaultParamHandler to update the members
Auxiliary Functions for the search of matching regions

std::vector< DoubleReal >::iterator	searchInScan_ (std::vector< DoubleReal >::iterator scan_begin, std::vector< DoubleReal >::iterator scan_end, DoubleReal current_mz)
template<typename InputSpectrumIterator , typename OutputPeakType >
void	optimizeRegions_ (InputSpectrumIterator &first, InputSpectrumIterator &last, MSExperiment< OutputPeakType > &ms_exp)
template<typename InputSpectrumIterator , typename OutputPeakType >
void	optimizeRegionsScanwise_ (InputSpectrumIterator &first, InputSpectrumIterator &last, MSExperiment< OutputPeakType > &ms_exp)
template<typename InputSpectrumIterator , typename OutputPeakType >
void	getRegionEndpoints_ (MSExperiment< OutputPeakType > &exp, InputSpectrumIterator &first, InputSpectrumIterator &last, Size iso_map_idx, DoubleReal noise_level, TwoDOptimization::Data &d)
	Get the indices of the first and last raw data point of this region.
void	findMatchingPeaks_ (std::multimap< DoubleReal, IsotopeCluster >::iterator &it, MSExperiment<> &ms_exp)
	Identify matching peak in a peak cluster.
Static Protected Member Functions
Functions provided to the gsl Levenberg-Marquardt

static Int	residual2D_ (const gsl_vector x, void params, gsl_vector *f)
	Function computing estimated signal and its deviation to the experimental signal*/.
static Int	jacobian2D_ (const gsl_vector x, void params, gsl_matrix *J)
	Function computing the Jacobian */.
static Int	evaluate2D_ (const gsl_vector x, void params, gsl_vector f, gsl_matrix J)
	Function that calls residual2D and jacobian2D*/.
Protected Attributes
std::multimap< DoubleReal, IsotopeCluster >	iso_map_
	stores the retention time of each isotopic cluster
std::multimap< DoubleReal, IsotopeCluster > ::const_iterator	curr_region_
	Pointer to the current region.
DoubleReal	max_peak_distance_
	upper bound for distance between two peaks belonging to the same region
DoubleReal	tolerance_mz_
	threshold for the difference in the peak position of two matching peaks
std::map< Int, std::vector < PeakIndex > >	matching_peaks_
	Indices of peaks in the adjacent scans matching peaks in the scan with no. ref_scan.
DoubleReal	eps_abs_
	Convergence Parameter: Maximal absolute error.
DoubleReal	eps_rel_
	Convergence Parameter: Maximal relative error.
UInt	max_iteration_
	Convergence Parameter: Maximal number of iterations.
bool	real_2D_
	Optimization considering all scans of a cluster or optimization of each scan separately.
OptimizationFunctions::PenaltyFactorsIntensity	penalties_
	Penalty factors for some parameters in the optimization.

Detailed Description

This class provides the two-dimensional optimization of the picked peak parameters.

Given the picked peaks, this class optimizes the peak parameters of each isotope pattern using a non-linear optimization. The peaks of adjacent scans are adjusted to achieve that a peak occuring in several scans has always the same m/z position. For the optimization the Levenberg-Marquardt algorithm provided from the GSL is used. The optimized parameters are the m/z values, the left and right width, which shall be equal for a peak in all scans, and the peaks' heights.

Todo:: Works only with defined types due to pointers to the data in the optimization namespace! Change that or remove templates (Alexandra)

Parameters of this class are:

Name	Type	Default	Description
delta_abs_error	float	9.99999974737875e-06	if the absolute error gets smaller than this value the fitting is stopped.
delta_rel_error	float	9.99999974737875e-06	if the relative error gets smaller than this value the fitting is stopped.
iterations	int	10	maximal number of iterations for the fitting step
penalties:position	float	0	If the position changes more than 0.2Da during the fitting it can be penalized
penalties:height	float	1	penalty term for the fitting of the intensity:If it gets negative during the fitting it can be penalized.
penalties:left_width	float	0	penalty term for the fitting of the left width:If the left width gets too broad or negative during the fitting it can be penalized.
penalties:right_width	float	0	penalty term for the fitting of the right width:If the right width gets too broad or negative during the fitting it can be penalized.
2d:tolerance_mz	float	2.2	mz tolerance for cluster construction
2d:max_peak_distance	float	1.2	maximal peak distance in mz in a cluster

Note:

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

Constructor & Destructor Documentation

TwoDOptimization ( )

Constructor.

TwoDOptimization ( const TwoDOptimization & opt )

Copy constructor.

virtual ~TwoDOptimization ( ) [inline, virtual]

Destructor.

Member Function Documentation

static Int evaluate2D_	(	const gsl_vector *	x,
		void *	params,
		gsl_vector *	f,
		gsl_matrix *	J
	)			`[static, protected]`

Function that calls residual2D and jacobian2D*/.

void findMatchingPeaks_	(	std::multimap< DoubleReal, IsotopeCluster >::iterator &	it,
		MSExperiment<> &	ms_exp
	)			`[protected]`

Identify matching peak in a peak cluster.

Referenced by TwoDOptimization::optimizeRegions_().

DoubleReal getMaxAbsError ( ) const [inline]

Non-mutable access to the maximal absolute error.

UInt getMaxIterations ( ) const [inline]

Non-mutable access to the maximal number of iterations.

DoubleReal getMaxPeakDistance ( ) const [inline]

Non-mutable access to the maximal peak distance in a cluster.

DoubleReal getMaxRelError ( ) const [inline]

Non-mutable access to the maximal relative error.

DoubleReal getMZTolerance ( ) const [inline]

Non-mutable access to the matching epsilon.

const OptimizationFunctions::PenaltyFactorsIntensity& getPenalties ( ) const [inline]

Non-mutable access to the minimal number of adjacent scans.

void getRegionEndpoints_	(	MSExperiment< OutputPeakType > &	exp,
		InputSpectrumIterator &	first,
		InputSpectrumIterator &	last,
		Size	iso_map_idx,
		DoubleReal	noise_level,
		TwoDOptimization::Data &	d
	)			`[protected]`

Get the indices of the first and last raw data point of this region.

References TwoDOptimization::iso_map_, MSExperiment< PeakT, ChromatogramPeakT >::RTBegin(), MSSpectrum< PeakT >::setRT(), and TwoDOptimization::Data::signal2D.

Referenced by TwoDOptimization::optimizeRegions_(), and TwoDOptimization::optimizeRegionsScanwise_().

static Int jacobian2D_	(	const gsl_vector *	x,
		void *	params,
		gsl_matrix *	J
	)			`[static, protected]`

Function computing the Jacobian */.

TwoDOptimization& operator= ( const TwoDOptimization & opt )

Assignment operator.

void optimize	(	InputSpectrumIterator	first,
		InputSpectrumIterator	last,
		MSExperiment< OutputPeakType > &	ms_exp,
		bool	real2D = `true`
	)

Find two dimensional peak clusters and optimize their peak parameters.

Note:

For the peak spectra, the following meta data arrays (see MSSpectrum) have to be present and have to be named just as listed here:

intensity (index:1)
leftWidth (index:3)
rightWidth (index:4)
peakShape (index:5)

Parameters:

	first	begin of the raw data spectra iterator range
	last	end of the raw data spectra interator range
	ms_exp	peak map corresponding to the raw data in the range from `first` to `last`
	real2D	flag if the optimization should be two dimensional or on each scan separately

Exceptions:

Exception::IllegalArgument

is thrown if required meta information from peak picking is missing (area, shape, left width, right width) or if the input data is invalid in some other way

References TwoDOptimization::curr_region_, Param::getValue(), TwoDOptimization::iso_map_, TwoDOptimization::max_peak_distance_, TwoDOptimization::optimizeRegions_(), TwoDOptimization::optimizeRegionsScanwise_(), DefaultParamHandler::param_, IsotopeCluster::peaks, TwoDOptimization::real_2D_, IsotopeCluster::scans, TwoDOptimization::searchInScan_(), and MSSpectrum< PeakT >::setRT().

void optimizeRegions_	(	InputSpectrumIterator &	first,
		InputSpectrumIterator &	last,
		MSExperiment< OutputPeakType > &	ms_exp
	)			`[protected]`

Performs 2D optimization of all regions

References TwoDOptimization::eps_abs_, TwoDOptimization::eps_rel_, TwoDOptimization::findMatchingPeaks_(), TwoDOptimization::getRegionEndpoints_(), TwoDOptimization::iso_map_, TwoDOptimization::Data::iso_map_iter, OpenMS::Constants::k, TwoDOptimization::Data::matching_peaks, TwoDOptimization::matching_peaks_, TwoDOptimization::max_iteration_, TwoDOptimization::Data::penalties, TwoDOptimization::penalties_, TwoDOptimization::Data::picked_peaks, TwoDOptimization::Data::raw_data_first, TwoDOptimization::Data::signal2D, and TwoDOptimization::Data::total_nr_peaks.

Referenced by TwoDOptimization::optimize().

void optimizeRegionsScanwise_	(	InputSpectrumIterator &	first,
		InputSpectrumIterator &	last,
		MSExperiment< OutputPeakType > &	ms_exp
	)			`[protected]`

static Int residual2D_	(	const gsl_vector *	x,
		void *	params,
		gsl_vector *	f
	)			`[static, protected]`

Function computing estimated signal and its deviation to the experimental signal*/.

std::vector<DoubleReal>::iterator searchInScan_	(	std::vector< DoubleReal >::iterator	scan_begin,
		std::vector< DoubleReal >::iterator	scan_end,
		DoubleReal	current_mz
	)			`[protected]`