A variant of QT clustering for the detection of feature groups. More...

#include <OpenMS/ANALYSIS/MAPMATCHING/QTClusterFinder.h>

Inheritance diagram for QTClusterFinder:

Collaboration diagram for QTClusterFinder:

Public Types
typedef std::unordered_map< std::pair< OpenMS::GridFeature , OpenMS::GridFeature >, double >	PairDistances
	Distances between pairs of grid features. More...

typedef std::unordered_map< const OpenMS::GridFeature *, std::unordered_set< Size > >	ElementMapping
	Map to store which grid features are next to which clusters (saves the clusters ids) More...

typedef boost::heap::fibonacci_heap< QTCluster >	Heap
	Heap to efficiently find the best clusters. More...

typedef HashGrid< OpenMS::GridFeature * >	Grid

Public Types inherited from ProgressLogger
enum	LogType { CMD , GUI , NONE }
	Possible log types. More...

Public Member Functions
	QTClusterFinder ()
	Constructor. More...

	~QTClusterFinder () override
	Destructor. More...

void	run (const std::vector< ConsensusMap > &input_maps, ConsensusMap &result_map) override
	Runs the algorithm on consensus maps. More...

void	run (const std::vector< FeatureMap > &input_maps, ConsensusMap &result_map)
	Runs the algorithm on feature maps. More...

Public Member Functions inherited from BaseGroupFinder
	BaseGroupFinder ()
	Default constructor. More...

	~BaseGroupFinder () override
	Destructor. More...

Public Member Functions inherited from DefaultParamHandler
	DefaultParamHandler (const String &name)
	Constructor with name that is displayed in error messages. More...

	DefaultParamHandler (const DefaultParamHandler &rhs)
	Copy constructor. More...

virtual	~DefaultParamHandler ()
	Destructor. More...

DefaultParamHandler &	operator= (const DefaultParamHandler &rhs)
	Assignment operator. More...

virtual bool	operator== (const DefaultParamHandler &rhs) const
	Equality operator. More...

void	setParameters (const Param &param)
	Sets the parameters. More...

const Param &	getParameters () const
	Non-mutable access to the parameters. More...

const Param &	getDefaults () const
	Non-mutable access to the default parameters. More...

const String &	getName () const
	Non-mutable access to the name. More...

void	setName (const String &name)
	Mutable access to the name. More...

const std::vector< String > &	getSubsections () const
	Non-mutable access to the registered subsections. More...

Public Member Functions inherited from ProgressLogger
	ProgressLogger ()
	Constructor. More...

virtual	~ProgressLogger ()
	Destructor. More...

	ProgressLogger (const ProgressLogger &other)
	Copy constructor. More...

ProgressLogger &	operator= (const ProgressLogger &other)
	Assignment Operator. More...

void	setLogType (LogType type) const
	Sets the progress log that should be used. The default type is NONE! More...

LogType	getLogType () const
	Returns the type of progress log being used. More...

void	startProgress (SignedSize begin, SignedSize end, const String &label) const
	Initializes the progress display. More...

void	setProgress (SignedSize value) const
	Sets the current progress. More...

void	endProgress () const
	Ends the progress display. More...

void	nextProgress () const
	increment progress by 1 (according to range begin-end) More...

Static Public Member Functions
static const String	getProductName ()
	Returns the name of the product. More...

static BaseGroupFinder *	create ()
	Returns an instance of this class. More...

Static Public Member Functions inherited from BaseGroupFinder
static void	registerChildren ()
	Register all derived classes here. More...

Static Public Member Functions inherited from DefaultParamHandler
static void	writeParametersToMetaValues (const Param &write_this, MetaInfoInterface &write_here, const String &key_prefix="")
	Writes all parameters to meta values. More...

Protected Types
enum	{ RT = Peak2D::RT , MZ = Peak2D::MZ }

Private Member Functions
double	getDistance_ (const OpenMS::GridFeature left, const OpenMS::GridFeature right)
	Calculates the distance between two grid features. More...

void	setParameters_ (double max_intensity, double max_mz)
	Sets algorithm parameters. More...

bool	makeConsensusFeature_ (Heap &cluster_heads, ConsensusFeature &feature, ElementMapping &element_mapping, const Grid &grid, const std::vector< Heap::handle_type > &handles)
	Extract the best cluster from cluster_heads and turn it into a consensus feature. More...

void	computeClustering_ (const Grid &grid, Heap &cluster_heads, std::vector< QTCluster::BulkData > &cluster_data, std::vector< Heap::handle_type > &handles, ElementMapping &element_mapping)
	Computes an initial QT clustering of the points in the hash grid. More...

void	removeFromElementMapping_ (const QTCluster &cluster, ElementMapping &element_mapping)
	Removes id of current top cluster in the heap from element mapping. More...

void	createConsensusFeature_ (ConsensusFeature &feature, const double quality, const QTCluster::Elements &elements)
	creates a consensus feature from the given elements More...

void	updateClustering_ (ElementMapping &element_mapping, const Grid &grid, const QTCluster::Elements &elements, Heap &cluster_heads, const std::vector< Heap::handle_type > &handles, Size best_id)
	update the clustering: More...

template<typename MapType >
void	run_ (const std::vector< MapType > &input_maps, ConsensusMap &result_map)
	Runs the algorithm on feature maps or consensus maps. More...

template<typename MapType >
void	run_internal_ (const std::vector< MapType > &input_maps, ConsensusMap &result_map, bool do_progress)
	Runs the algorithm on feature maps or consensus maps (internal) More...

void	addClusterElements_ (const Grid &grid, QTCluster &cluster)
	Adds elements to the cluster based on the elements hashed in the grid. More...

bool	distIsOutlier_ (double dist, double rt)
	Looks up the matching bin for `rt` in bin_tolerances_ and checks if `dist` is in the allowed range. More...

Private Attributes
Size	num_maps_
	Number of input maps. More...

bool	use_IDs_
	Consider peptide identifications for grouping? More...

Size	min_nr_diffs_per_bin_
	Min. nr. of differences from matched IDs requested to calculate a linking tolerance per RT bin. More...

double	min_score_
	Min. score for an ID to be considered for tolerance estimation. More...

double	noID_penalty_

double	max_diff_rt_
	Maximum RT difference. More...

double	max_diff_mz_
	Maximum m/z difference. More...

int	nr_partitions_
	Maximum m/z difference. More...

FeatureDistance	feature_distance_
	Feature distance functor. More...

std::unordered_set< const OpenMS::GridFeature * >	already_used_
	Set of features already used. More...

std::map< double, double >	bin_tolerances_

Additional Inherited Members
Protected Member Functions inherited from BaseGroupFinder
void	checkIds_ (const std::vector< ConsensusMap > &maps) const
	Checks if all file descriptions have disjoint map identifiers. More...

Protected Member Functions inherited from DefaultParamHandler
virtual void	updateMembers_ ()
	This method is used to update extra member variables at the end of the setParameters() method. More...

void	defaultsToParam_ ()
	Updates the parameters after the defaults have been set in the constructor. More...

Static Protected Member Functions inherited from ProgressLogger
static String	logTypeToFactoryName_ (LogType type)
	Return the name of the factory product used for this log type. More...

Protected Attributes inherited from DefaultParamHandler
Param	param_
	Container for current parameters. More...

Param	defaults_
	Container for default parameters. This member should be filled in the constructor of derived classes! More...

std::vector< String >	subsections_
	Container for registered subsections. This member should be filled in the constructor of derived classes! More...

String	error_name_
	Name that is displayed in error messages during the parameter checking. More...

bool	check_defaults_
	If this member is set to false no checking if parameters in done;. More...

bool	warn_empty_defaults_
	If this member is set to false no warning is emitted when defaults are empty;. More...

Protected Attributes inherited from ProgressLogger
LogType	type_

time_t	last_invoke_

ProgressLoggerImpl *	current_logger_

Static Protected Attributes inherited from ProgressLogger
static int	recursion_depth_

Detailed Description

A variant of QT clustering for the detection of feature groups.

The algorithm accumulates all features from all input maps, then applies a variant of QT clustering to find groups of corresponding features. In more detail, every feature from every input map is considered as a potential cluster center. For every center, its nearest neighbors from the other input maps are detected and added to the potential cluster. Iteratively, the cluster with the highest quality is extracted and the clustering is updated.

Properties affecting the grouping

To be included in a particular cluster, a feature has to fulfill the following conditions:

differences in RT and m/z from the cluster center must be below user-defined thresholds (distance_RT:max_difference and distance_MZ:max_difference),

the charge state must match that of the cluster center (unless ignore_charge is set),

if use_identifications is set and both the feature and the cluster center are annotated with peptide identifications, the identifications have to match.

Every cluster contains at most one feature from each input map - namely the feature closest to the cluster center that meets the criteria and does not belong to a better cluster.

The notion of "closeness" for features is defined by the distance function implemented in FeatureDistance, the parameters of which can be set by the user.

The quality of a cluster is computed from the number of elements in it and their distances to the cluster center. For more details see QTCluster.

Optimization

This algorithm includes a number of optimizations to reduce run-time:

two-dimensional hashing of features,
a look-up table for feature distances,
a variant of QT clustering that requires only one round of clustering.

See also: FeatureGroupingAlgorithmQT

Parameters of this class are:

Name	Type	Default	Restrictions	Description
use_identifications	string	false	true, false	Never link features that are annotated with different peptides (only the best hit per peptide identification is taken into account).
nr_partitions	int	100	min: 1	How many partitions in m/z space should be used for the algorithm (more partitions means faster runtime and more memory efficient execution).
min_nr_diffs_per_bin	int	50	min: 5	If IDs are used: How many differences from matching IDs should be used to calculate a linking tolerance for unIDed features in an RT region. RT regions will be extended until that number is reached.
min_IDscore_forTolCalc	float	1.0		If IDs are used: What is the minimum score of an ID to assume a reliable match for tolerance calculation. Check your current score type!
noID_penalty	float	0.0	min: 0.0 max: 1.0	If IDs are used: For the normalized distances, how high should the penalty for missing IDs be? 0 = no bias, 1 = IDs inside the max tolerances always preferred (even if much further away).
ignore_charge	string	false	true, false	false [default]: pairing requires equal charge state (or at least one unknown charge '0'); true: Pairing irrespective of charge state
ignore_adduct	string	true	true, false	true [default]: pairing requires equal adducts (or at least one without adduct annotation); true: Pairing irrespective of adducts
distance_RT:max_difference	float	100.0	min: 0.0	Never pair features with a larger RT distance (in seconds).
distance_RT:exponent	float	1.0	min: 0.0	Normalized RT differences ([0-1], relative to 'max_difference') are raised to this power (using 1 or 2 will be fast, everything else is REALLY slow)
distance_RT:weight	float	1.0	min: 0.0	Final RT distances are weighted by this factor
distance_MZ:max_difference	float	0.3	min: 0.0	Never pair features with larger m/z distance (unit defined by 'unit')
distance_MZ:unit	string	Da	Da, ppm	Unit of the 'max_difference' parameter
distance_MZ:exponent	float	2.0	min: 0.0	Normalized ([0-1], relative to 'max_difference') m/z differences are raised to this power (using 1 or 2 will be fast, everything else is REALLY slow)
distance_MZ:weight	float	1.0	min: 0.0	Final m/z distances are weighted by this factor
distance_intensity:exponent	float	1.0	min: 0.0	Differences in relative intensity ([0-1]) are raised to this power (using 1 or 2 will be fast, everything else is REALLY slow)
distance_intensity:weight	float	0.0	min: 0.0	Final intensity distances are weighted by this factor
distance_intensity:log_transform	string	disabled	enabled, disabled	Log-transform intensities? If disabled, d = \|int_f2 - int_f1\| / int_max. If enabled, d = \|log(int_f2 + 1) - log(int_f1 + 1)\| / log(int_max + 1))

Note:

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

Member Typedef Documentation

◆ ElementMapping

typedef std::unordered_map< const OpenMS::GridFeature*, std::unordered_set<Size> > ElementMapping

Map to store which grid features are next to which clusters (saves the clusters ids)

◆ Grid

typedef HashGrid<OpenMS::GridFeature*> Grid

◆ Heap

typedef boost::heap::fibonacci_heap<QTCluster> Heap

Heap to efficiently find the best clusters.

◆ PairDistances

typedef std::unordered_map< std::pair<OpenMS::GridFeature*, OpenMS::GridFeature*>, double> PairDistances

Distances between pairs of grid features.

Member Enumeration Documentation

◆ anonymous enum

anonymous enum

protected

Enumerator
RT
MZ

Constructor & Destructor Documentation

◆ QTClusterFinder()

QTClusterFinder ( )

Constructor.

◆ ~QTClusterFinder()

~QTClusterFinder ( )

override

Destructor.

Member Function Documentation

◆ addClusterElements_()

void addClusterElements_	(	const Grid &	grid,
		QTCluster &	cluster
	)

private

Adds elements to the cluster based on the elements hashed in the grid.

Parameters

grid	the grid is used to find neighboring features the cluster
cluster	cluster to which the new elements are added

◆ computeClustering_()

void computeClustering_	(	const Grid &	grid,
		Heap &	cluster_heads,
		std::vector< QTCluster::BulkData > &	cluster_data,
		std::vector< Heap::handle_type > &	handles,
		ElementMapping &	element_mapping
	)

private

Computes an initial QT clustering of the points in the hash grid.

Parameters

grid	the grid is used to find new features for clusters that have to be updated
cluster_heads	the heap where the QTClusters are inserted
cluster_data	vector where the BulkData objects of the clusters are stored
handles	vector where handles of the inserted clusters are stored
element_mapping	the element mapping where all the clusters get registered for their features

◆ create()

static BaseGroupFinder* create ( )

inlinestatic

Returns an instance of this class.

◆ createConsensusFeature_()

void createConsensusFeature_	(	ConsensusFeature &	feature,
		const double	quality,
		const QTCluster::Elements &	elements
	)

private

creates a consensus feature from the given elements

Parameters

[out]	feature	The resulting consensus feature which is constructed here
	quality	the quality of the new consensus feature
	elements	the original features that from the new consensus feature

Note: the features from elements are registered in the already_used_ member of this class

◆ distIsOutlier_()

bool distIsOutlier_	(	double	dist,
		double	rt
	)

private

Looks up the matching bin for rt in bin_tolerances_ and checks if dist is in the allowed range.

◆ getDistance_()

double getDistance_	(	const OpenMS::GridFeature *	left,
		const OpenMS::GridFeature *	right
	)

private

Calculates the distance between two grid features.

◆ getProductName()

static const String getProductName ( )

inlinestatic

Returns the name of the product.

◆ makeConsensusFeature_()

bool makeConsensusFeature_	(	Heap &	cluster_heads,
		ConsensusFeature &	feature,
		ElementMapping &	element_mapping,
		const Grid &	grid,
		const std::vector< Heap::handle_type > &	handles
	)

private

Extract the best cluster from cluster_heads and turn it into a consensus feature.

Parameters

	cluster_heads	the heap where the clusters are stored, must not be empty
[out]	feature	The resulting consensus feature which is constructed here
	element_mapping	the element mapping is used to update clusters when features are removed
	grid	the grid is used to find new features for clusters that have to be updated
	handles	used to access clusters if we know their id from the element mapping

Returns: bool whether a consensus feature was made or not

◆ removeFromElementMapping_()

void removeFromElementMapping_	(	const QTCluster &	cluster,
		ElementMapping &	element_mapping
	)

private

Removes id of current top cluster in the heap from element mapping.

Parameters

cluster	the current top cluster in the heap which id is removed
element_mapping	the element mapping from which the ids are removed

◆ run() [1/2]

void run	(	const std::vector< ConsensusMap > &	input_maps,
		ConsensusMap &	result_map
	)

overridevirtual

Runs the algorithm on consensus maps.

Precondition: The data ranges of the input maps have to be up-to-date (use ConsensusMap::updateRanges).

Exceptions

Exception::IllegalArgument is thrown if the input data is not valid.

Implements BaseGroupFinder.

◆ run() [2/2]

void run	(	const std::vector< FeatureMap > &	input_maps,
		ConsensusMap &	result_map
	)

Runs the algorithm on feature maps.

Precondition: The data ranges of the input maps have to be up-to-date (use FeatureMap::updateRanges).

Exceptions

Exception::IllegalArgument is thrown if the input data is not valid.

◆ run_()

void run_	(	const std::vector< MapType > &	input_maps,
		ConsensusMap &	result_map
	)

private

Runs the algorithm on feature maps or consensus maps.

◆ run_internal_()

void run_internal_	(	const std::vector< MapType > &	input_maps,
		ConsensusMap &	result_map,
		bool	do_progress
	)

private

Runs the algorithm on feature maps or consensus maps (internal)

◆ setParameters_()

void setParameters_	(	double	max_intensity,
		double	max_mz
	)

private

Sets algorithm parameters.

◆ updateClustering_()

void updateClustering_	(	ElementMapping &	element_mapping,
		const Grid &	grid,
		const QTCluster::Elements &	elements,
		Heap &	cluster_heads,
		const std::vector< Heap::handle_type > &	handles,
		Size	best_id
	)

private

update the clustering:

1. remove current best cluster from the heap 2. update all clusters accordingly by removing neighbors used by the current best 3. invalidate clusters whose center has been used by the current best

Parameters

element_mapping	the element mapping is used to update clusters and updated itself
grid	the grid is used to find new features for clusters that have to be updated
cluster_heads	the heap is updated for changing clusters and popped in the end
elements	the features that now have to be removed from other clusters than the current best
handles	used to access clusters if we know their id from the element mapping
best_id	id of the current best cluster, will be removed from the element mapping

Note: The feature from elements are not deleted from the element mapping. After this function is called we don't have any cluster with those features left and therefore don't have to delete them.

Member Data Documentation

◆ already_used_

std::unordered_set<const OpenMS::GridFeature*> already_used_

private

Set of features already used.

◆ bin_tolerances_

std::map<double, double> bin_tolerances_

private

Map of median RTs to allowed linking tolerances (on the same RT scale) for unIDed features. This should be interpreted as bins from the current median RT to the next.

◆ feature_distance_

FeatureDistance feature_distance_

private

Feature distance functor.

◆ max_diff_mz_

double max_diff_mz_

private

Maximum m/z difference.

◆ max_diff_rt_

double max_diff_rt_

private

Maximum RT difference.

◆ min_nr_diffs_per_bin_

Size min_nr_diffs_per_bin_

private

Min. nr. of differences from matched IDs requested to calculate a linking tolerance per RT bin.

◆ min_score_

double min_score_

private

Min. score for an ID to be considered for tolerance estimation.

◆ noID_penalty_

double noID_penalty_

private

Distance penalty for unidentified features when finding best neighbor per map and for cluster quality calculation and therefore the order in which they are popped from the heap. Since distances are normalized, a penalty of 1.0 will always prefer IDed features.

◆ nr_partitions_

int nr_partitions_

private

Maximum m/z difference.

◆ num_maps_

Size num_maps_

private

Number of input maps.

◆ use_IDs_

bool use_IDs_

private

Consider peptide identifications for grouping?

Public Types

Public Member Functions

Static Public Member Functions

Protected Types

Private Member Functions

Private Attributes

Additional Inherited Members

Detailed Description

Member Typedef Documentation

◆ ElementMapping

◆ Grid

◆ Heap

◆ PairDistances

Member Enumeration Documentation

◆ anonymous enum

Constructor & Destructor Documentation

◆ QTClusterFinder()

◆ ~QTClusterFinder()

Member Function Documentation

◆ addClusterElements_()

◆ computeClustering_()

◆ create()

◆ createConsensusFeature_()

◆ distIsOutlier_()

◆ getDistance_()

◆ getProductName()

◆ makeConsensusFeature_()

◆ removeFromElementMapping_()

◆ run() [1/2]

◆ run() [2/2]

◆ run_()

◆ run_internal_()

◆ setParameters_()

◆ updateClustering_()

Member Data Documentation

◆ already_used_

◆ bin_tolerances_

◆ feature_distance_

◆ max_diff_mz_

◆ max_diff_rt_

◆ min_nr_diffs_per_bin_

◆ min_score_

◆ noID_penalty_

◆ nr_partitions_

◆ num_maps_

◆ use_IDs_