Collection of functions for filtering peptide and protein identifications. More...

#include <OpenMS/FILTERING/ID/IDFilter.h>

Classes
struct	DigestionFilter
	Is peptide evidence digestion product of some protein. More...

struct	GetMatchingItems
	Builds a map index of data that have a String index to find matches and return the objects. More...

struct	HasDecoyAnnotation
	Is this a decoy hit? More...

struct	HasGoodScore
	Is the score of this hit at least as good as the given value? More...

struct	HasMatchingAccession
	Given a list of protein accessions, do any occur in the annotation(s) of this hit? More...

struct	HasMaxMetaValue
	Does a meta value of this hit have at most the given value? More...

struct	HasMaxRank
	Is the rank of this hit below or at the given cut-off? More...

struct	HasMetaValue
	Is a meta value with given key and value set on this hit? More...

struct	HasNoHits
	Is the list of hits of this peptide/protein ID empty? More...

class	PeptideDigestionFilter
	Filter Peptide Hit by its digestion product. More...

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

virtual	~IDFilter ()
	Destructor. More...

Static Public Member Functions
Higher-order filter functions
Functions for filtering a container based on a predicate
template<class Container , class Predicate >
static void	removeMatchingItems (Container &items, const Predicate &pred)
	Remove items that satisfy a condition from a container (e.g. vector) More...

template<class Container , class Predicate >
static void	keepMatchingItems (Container &items, const Predicate &pred)
	Keep items that satisfy a condition in a container (e.g. vector), removing all others. More...

Helper functions
template<class IdentificationType >
static Size	countHits (const std::vector< IdentificationType > &ids)
	Returns the total number of peptide/protein hits in a vector of peptide/protein identifications. More...

template<class IdentificationType >
static bool	getBestHit (const std::vector< IdentificationType > &identifications, bool assume_sorted, typename IdentificationType::HitType &best_hit)
	Finds the best-scoring hit in a vector of peptide or protein identifications. More...

static void	extractPeptideSequences (const std::vector< PeptideIdentification > &peptides, std::set< String > &sequences, bool ignore_mods=false)
	Extracts all unique peptide sequences from a list of peptide IDs. More...

template<class EvidenceFilter >
static void	FilterPeptideEvidences (EvidenceFilter &filter, std::vector< PeptideIdentification > &peptides)
	remove peptide evidences based on a filter More...

Clean-up functions
template<class IdentificationType >
static void	updateHitRanks (std::vector< IdentificationType > &ids)
	Updates the hit ranks on all peptide or protein IDs. More...

static void	removeUnreferencedProteins (std::vector< ProteinIdentification > &proteins, const std::vector< PeptideIdentification > &peptides)
	Removes protein hits from `proteins` that are not referenced by a peptide in `peptides`. More...

static void	updateProteinReferences (std::vector< PeptideIdentification > &peptides, const std::vector< ProteinIdentification > &proteins, bool remove_peptides_without_reference=false)
	Removes references to missing proteins. More...

static bool	updateProteinGroups (std::vector< ProteinIdentification::ProteinGroup > &groups, const std::vector< ProteinHit > &hits)
	Update protein groups after protein hits were filtered. More...

Filter functions for peptide or protein IDs
template<class IdentificationType >
static void	removeEmptyIdentifications (std::vector< IdentificationType > &ids)
	Removes peptide or protein identifications that have no hits in them. More...

template<class IdentificationType >
static void	filterHitsByScore (std::vector< IdentificationType > &ids, double threshold_score)
	Filters peptide or protein identifications according to the score of the hits. More...

template<class IdentificationType >
static void	filterHitsBySignificance (std::vector< IdentificationType > &ids, double threshold_fraction=1.0)
	Filters peptide or protein identifications according to the significance threshold of the hits. More...

template<class IdentificationType >
static void	keepNBestHits (std::vector< IdentificationType > &ids, Size n)
	Filters peptide or protein identifications according to the score of the hits, keeping the `n` best hits per ID. More...

template<class IdentificationType >
static void	filterHitsByRank (std::vector< IdentificationType > &ids, Size min_rank, Size max_rank)
	Filters peptide or protein identifications according to the ranking of the hits. More...

template<class IdentificationType >
static void	removeDecoyHits (std::vector< IdentificationType > &ids)
	Removes hits annotated as decoys from peptide or protein identifications. More...

template<class IdentificationType >
static void	removeHitsMatchingProteins (std::vector< IdentificationType > &ids, const std::set< String > accessions)
	Filters peptide or protein identifications according to the given proteins (negative). More...

template<class IdentificationType >
static void	keepHitsMatchingProteins (std::vector< IdentificationType > &ids, const std::set< String > accessions)
	Filters peptide or protein identifications according to the given proteins (positive). More...

Filter functions for peptide IDs only
static void	keepBestPeptideHits (std::vector< PeptideIdentification > &peptides, bool strict=false)
	Filters peptide identifications keeping only the single best-scoring hit per ID. More...

static void	filterPeptidesByLength (std::vector< PeptideIdentification > &peptides, Size min_length, Size max_length=UINT_MAX)
	Filters peptide identifications according to peptide sequence length. More...

static void	filterPeptidesByCharge (std::vector< PeptideIdentification > &peptides, Int min_charge, Int max_charge)
	Filters peptide identifications according to charge state. More...

static void	filterPeptidesByRT (std::vector< PeptideIdentification > &peptides, double min_rt, double max_rt)
	Filters peptide identifications by precursor RT, keeping only IDs in the given range. More...

static void	filterPeptidesByMZ (std::vector< PeptideIdentification > &peptides, double min_mz, double max_mz)
	Filters peptide identifications by precursor m/z, keeping only IDs in the given range. More...

static void	filterPeptidesByMZError (std::vector< PeptideIdentification > &peptides, double mass_error, bool unit_ppm)
	Filter peptide identifications according to mass deviation. More...

template<class Filter >
static void	filterPeptideEvidences (Filter &filter, std::vector< PeptideIdentification > &peptides)
	Digest a collection of proteins and filter PeptideEvidences based on specificity PeptideEvidences of peptides are removed if the digest of a protein did not produce the peptide sequence. More...

static void	filterPeptidesByRTPredictPValue (std::vector< PeptideIdentification > &peptides, const String &metavalue_key, double threshold=0.05)
	Filters peptide identifications according to p-values from RTPredict. More...

static void	removePeptidesWithMatchingModifications (std::vector< PeptideIdentification > &peptides, const std::set< String > &modifications)
	Removes all peptide hits that have at least one of the given modifications. More...

static void	keepPeptidesWithMatchingModifications (std::vector< PeptideIdentification > &peptides, const std::set< String > &modifications)
	Keeps only peptide hits that have at least one of the given modifications. More...

static void	removePeptidesWithMatchingSequences (std::vector< PeptideIdentification > &peptides, const std::vector< PeptideIdentification > &bad_peptides, bool ignore_mods=false)
	Removes all peptide hits with a sequence that matches one in `bad_peptides`. More...

static void	keepPeptidesWithMatchingSequences (std::vector< PeptideIdentification > &peptides, const std::vector< PeptideIdentification > &good_peptides, bool ignore_mods=false)
	Removes all peptide hits with a sequence that does not match one in `good_peptides`. More...

static void	keepUniquePeptidesPerProtein (std::vector< PeptideIdentification > &peptides)
	Removes all peptides that are not annotated as unique for a protein (by PeptideIndexer) More...

static void	removeDuplicatePeptideHits (std::vector< PeptideIdentification > &peptides, bool seq_only=false)
	Removes duplicate peptide hits from each peptide identification, keeping only unique hits (per ID). More...

Filter functions for MS/MS experiments
static void	filterHitsByScore (PeakMap &experiment, double peptide_threshold_score, double protein_threshold_score)
	Filters an MS/MS experiment according to score thresholds. More...

static void	filterHitsBySignificance (PeakMap &experiment, double peptide_threshold_fraction, double protein_threshold_fraction)
	Filters an MS/MS experiment according to fractions of the significance thresholds. More...

static void	keepNBestHits (PeakMap &experiment, Size n)
	Filters an MS/MS experiment by keeping the N best peptide hits for every spectrum. More...

static void	keepHitsMatchingProteins (PeakMap &experiment, const std::vector< FASTAFile::FASTAEntry > &proteins)
	Filters an MS/MS experiment according to the given proteins. More...

Detailed Description

Collection of functions for filtering peptide and protein identifications.

This class provides functions for filtering collections of peptide or protein identifications according to various criteria. It also contains helper functions and classes (functors that implement predicates) that are used in this context.

The filter functions modify their inputs, rather than creating filtered copies.

Most filters work on the hit level, i.e. they remove peptide or protein hits from peptide or protein identifications (IDs). A few filters work on the ID level instead, i.e. they remove peptide or protein IDs from vectors thereof. Independent of this, the inputs for all filter functions are vectors of IDs, because the data most often comes in this form. This design also allows many helper objects to be set up only once per vector, rather than once per ID.

The filter functions for vectors of peptide/protein IDs do not include clean-up steps (e.g. removal of IDs without hits, reassignment of hit ranks, ...). They only carry out their specific filtering operations. This is so filters can be chained without having to repeat clean-up operations. The group of clean-up functions provides helpers that are useful to ensure data integrity after filters have been applied, but it is up to the individual developer to use them when necessary.

The filter functions for MS/MS experiments do include clean-up steps, because they filter peptide and protein IDs in conjunction and potential contradictions between the two must be eliminated.

Constructor & Destructor Documentation

◆ IDFilter()

IDFilter ( )

Constructor.

◆ ~IDFilter()

virtual ~IDFilter ( )

virtual

Destructor.

Member Function Documentation

◆ countHits()

static Size countHits ( const std::vector< IdentificationType > & ids )

inlinestatic

Returns the total number of peptide/protein hits in a vector of peptide/protein identifications.

◆ extractPeptideSequences()

static void extractPeptideSequences	(	const std::vector< PeptideIdentification > &	peptides,
		std::set< String > &	sequences,
		bool	ignore_mods = `false`
	)

static

Extracts all unique peptide sequences from a list of peptide IDs.

Parameters

peptides	Input
sequences	Output
ignore_mods	Extract sequences without modifications?

◆ filterHitsByRank()

static void filterHitsByRank	(	std::vector< IdentificationType > &	ids,
		Size	min_rank,
		Size	max_rank
	)

inlinestatic

Filters peptide or protein identifications according to the ranking of the hits.

The hits between min_rank and max_rank (both inclusive) in each ID are kept. Counting starts at 1, i.e. the best (highest/lowest scoring) hit has rank 1. The ranks are (re-)computed before filtering. max_rank is ignored if it is smaller than min_rank.

Note that there may be several hits with the same rank in a peptide or protein ID (if the scores are the same).

This method is useful if a range of higher hits is needed for decoy fairness analysis.

Note: The ranks of the hits may be invalidated.

◆ filterHitsByScore() [1/2]

static void filterHitsByScore	(	std::vector< IdentificationType > &	ids,
		double	threshold_score
	)

inlinestatic

Filters peptide or protein identifications according to the score of the hits.

Only peptide/protein hits with a score at least as good as threshold_score are kept. Score orientation (are higher scores better?) is taken into account.

◆ filterHitsByScore() [2/2]

static void filterHitsByScore	(	PeakMap &	experiment,
		double	peptide_threshold_score,
		double	protein_threshold_score
	)

inlinestatic

Filters an MS/MS experiment according to score thresholds.

References MSExperiment::begin(), MSExperiment::end(), and ExperimentalSettings::getProteinIdentifications().

◆ filterHitsBySignificance() [1/2]

static void filterHitsBySignificance	(	std::vector< IdentificationType > &	ids,
		double	threshold_fraction = `1.0`
	)

inlinestatic

Filters peptide or protein identifications according to the significance threshold of the hits.

Only peptide/protein hits which reach a score above (or below, depending on score orientation) threshold_fraction * significance_threshold (as stored in the ID) are kept.

◆ filterHitsBySignificance() [2/2]

static void filterHitsBySignificance	(	PeakMap &	experiment,
		double	peptide_threshold_fraction,
		double	protein_threshold_fraction
	)

inlinestatic

Filters an MS/MS experiment according to fractions of the significance thresholds.

References MSExperiment::begin(), MSExperiment::end(), and ExperimentalSettings::getProteinIdentifications().

◆ FilterPeptideEvidences()

static void FilterPeptideEvidences	(	EvidenceFilter &	filter,
		std::vector< PeptideIdentification > &	peptides
	)

inlinestatic

remove peptide evidences based on a filter

Parameters

filter	filter function that overloads ()(PeptideEvidence&) operator
peptides	a collection of peptide evidences

◆ filterPeptideEvidences()

static void filterPeptideEvidences	(	Filter &	filter,
		std::vector< PeptideIdentification > &	peptides
	)

static

Digest a collection of proteins and filter PeptideEvidences based on specificity PeptideEvidences of peptides are removed if the digest of a protein did not produce the peptide sequence.

Parameters

filter	filter function on PeptideEvidence level
peptides	PeptideIdentification that will be scanned and filtered

◆ filterPeptidesByCharge()

static void filterPeptidesByCharge	(	std::vector< PeptideIdentification > &	peptides,
		Int	min_charge,
		Int	max_charge
	)

static

Filters peptide identifications according to charge state.

Only peptide hits with a charge state between min_charge and max_charge (both inclusive) are kept. max_charge is ignored if it is smaller than min_charge.

Note: The ranks of the hits may be invalidated.

◆ filterPeptidesByLength()

static void filterPeptidesByLength	(	std::vector< PeptideIdentification > &	peptides,
		Size	min_length,
		Size	max_length = `UINT_MAX`
	)

static

Filters peptide identifications according to peptide sequence length.

Only peptide hits with a sequence length between min_length and max_length (both inclusive) are kept. max_length is ignored if it is smaller than min_length.

Note: The ranks of the hits may be invalidated.

◆ filterPeptidesByMZ()

static void filterPeptidesByMZ	(	std::vector< PeptideIdentification > &	peptides,
		double	min_mz,
		double	max_mz
	)

static

Filters peptide identifications by precursor m/z, keeping only IDs in the given range.

◆ filterPeptidesByMZError()

static void filterPeptidesByMZError	(	std::vector< PeptideIdentification > &	peptides,
		double	mass_error,
		bool	unit_ppm
	)

static

Filter peptide identifications according to mass deviation.

Only peptide hits with a low mass deviation (between theoretical peptide mass and precursor mass) are kept.

Parameters

identification	Input/output
mass_error	Threshold for the mass deviation
unit_ppm	Is `mass_error` given in PPM?

Note: The ranks of the hits may be invalidated.

◆ filterPeptidesByRT()

static void filterPeptidesByRT	(	std::vector< PeptideIdentification > &	peptides,
		double	min_rt,
		double	max_rt
	)

static

Filters peptide identifications by precursor RT, keeping only IDs in the given range.

◆ filterPeptidesByRTPredictPValue()

static void filterPeptidesByRTPredictPValue	(	std::vector< PeptideIdentification > &	peptides,
		const String &	metavalue_key,
		double	threshold = `0.05`
	)

static

Filters peptide identifications according to p-values from RTPredict.

Filters the peptide hits by the probability (p-value) of a correct peptide identification having a deviation between observed and predicted RT equal to or greater than allowed.

Parameters

peptides	Input/output
metavalue_key	Name of the meta value that holds the p-value: "predicted_RT_p_value" or "predicted_RT_p_value_first_dim"
threshold	P-value threshold

Note: The ranks of the hits may be invalidated.

◆ getBestHit()

static bool getBestHit	(	const std::vector< IdentificationType > &	identifications,
		bool	assume_sorted,
		typename IdentificationType::HitType &	best_hit
	)

inlinestatic

Finds the best-scoring hit in a vector of peptide or protein identifications.

If there are several hits with the best score, the first one is taken.

Parameters

identifications	Vector of peptide or protein IDs, each containing one or more (peptide/protein) hits
assume_sorted	Are hits sorted by score (best score first) already? This allows for faster query, since only the first hit needs to be looked at

Exception::InvalidValue if the IDs have different score types (i.e. scores cannot be compared)

Returns: true if a hit was present, false otherwise

◆ keepBestPeptideHits()

static void keepBestPeptideHits	(	std::vector< PeptideIdentification > &	peptides,
		bool	strict = `false`
	)

static

Filters peptide identifications keeping only the single best-scoring hit per ID.

Parameters

peptides	Input/output
strict	If set, keep the best hit only if its score is unique - i.e. ties are not allowed. (Otherwise all hits with the best score is kept.)

◆ keepHitsMatchingProteins() [1/2]

static void keepHitsMatchingProteins	(	std::vector< IdentificationType > &	ids,
		const std::set< String >	accessions
	)

inlinestatic

Filters peptide or protein identifications according to the given proteins (positive).

Hits with no matching protein accession in accessions are removed.

Note: The ranks of the hits may be invalidated.

◆ keepHitsMatchingProteins() [2/2]

static void keepHitsMatchingProteins	(	PeakMap &	experiment,
		const std::vector< FASTAFile::FASTAEntry > &	proteins
	)

inlinestatic

Filters an MS/MS experiment according to the given proteins.

References MSExperiment::begin(), MSExperiment::end(), and ExperimentalSettings::getProteinIdentifications().

◆ keepMatchingItems()

static void keepMatchingItems	(	Container &	items,
		const Predicate &	pred
	)

inlinestatic

Keep items that satisfy a condition in a container (e.g. vector), removing all others.

◆ keepNBestHits() [1/2]

static void keepNBestHits	(	std::vector< IdentificationType > &	ids,
		Size	n
	)

inlinestatic

Filters peptide or protein identifications according to the score of the hits, keeping the n best hits per ID.

The score orientation (are higher scores better?) is taken into account.

◆ keepNBestHits() [2/2]

static void keepNBestHits	(	PeakMap &	experiment,
		Size	n
	)

inlinestatic

Filters an MS/MS experiment by keeping the N best peptide hits for every spectrum.

References MSExperiment::begin(), MSExperiment::end(), and ExperimentalSettings::getProteinIdentifications().

◆ keepPeptidesWithMatchingModifications()

static void keepPeptidesWithMatchingModifications	(	std::vector< PeptideIdentification > &	peptides,
		const std::set< String > &	modifications
	)

static

Keeps only peptide hits that have at least one of the given modifications.

◆ keepPeptidesWithMatchingSequences()

static void keepPeptidesWithMatchingSequences	(	std::vector< PeptideIdentification > &	peptides,
		const std::vector< PeptideIdentification > &	good_peptides,
		bool	ignore_mods = `false`
	)

static

Removes all peptide hits with a sequence that does not match one in good_peptides.

If ignore_mods is set, unmodified sequences are generated and compared to the given ones.

Note: The ranks of the hits may be invalidated.

◆ keepUniquePeptidesPerProtein()

static void keepUniquePeptidesPerProtein ( std::vector< PeptideIdentification > & peptides )

static

Removes all peptides that are not annotated as unique for a protein (by PeptideIndexer)

◆ removeDecoyHits()

static void removeDecoyHits ( std::vector< IdentificationType > & ids )

inlinestatic

Removes hits annotated as decoys from peptide or protein identifications.

Checks for meta values named "target_decoy" and "isDecoy", and removes protein/peptide hits if the values are "decoy" and "true", respectively.

Note: The ranks of the hits may be invalidated.

◆ removeDuplicatePeptideHits()

static void removeDuplicatePeptideHits	(	std::vector< PeptideIdentification > &	peptides,
		bool	seq_only = `false`
	)

static

Removes duplicate peptide hits from each peptide identification, keeping only unique hits (per ID).

By default, hits are considered duplicated if they compare as equal using PeptideHit::operator==. However, if seq_only is set, only the sequences (incl. modifications) are compared. In both cases, the first occurrence of each hit in a peptide ID is kept, later ones are removed.

◆ removeEmptyIdentifications()

static void removeEmptyIdentifications ( std::vector< IdentificationType > & ids )

inlinestatic

Removes peptide or protein identifications that have no hits in them.

◆ removeHitsMatchingProteins()

static void removeHitsMatchingProteins	(	std::vector< IdentificationType > &	ids,
		const std::set< String >	accessions
	)

inlinestatic

Filters peptide or protein identifications according to the given proteins (negative).

Hits with a matching protein accession in accessions are removed.

Note: The ranks of the hits may be invalidated.

◆ removeMatchingItems()

static void removeMatchingItems	(	Container &	items,
		const Predicate &	pred
	)

inlinestatic

Remove items that satisfy a condition from a container (e.g. vector)

◆ removePeptidesWithMatchingModifications()

static void removePeptidesWithMatchingModifications	(	std::vector< PeptideIdentification > &	peptides,
		const std::set< String > &	modifications
	)

static

Removes all peptide hits that have at least one of the given modifications.

◆ removePeptidesWithMatchingSequences()

static void removePeptidesWithMatchingSequences	(	std::vector< PeptideIdentification > &	peptides,
		const std::vector< PeptideIdentification > &	bad_peptides,
		bool	ignore_mods = `false`
	)

static

Removes all peptide hits with a sequence that matches one in bad_peptides.

If ignore_mods is set, unmodified sequences are generated and compared to the given ones.

Note: The ranks of the hits may be invalidated.

◆ removeUnreferencedProteins()

static void removeUnreferencedProteins	(	std::vector< ProteinIdentification > &	proteins,
		const std::vector< PeptideIdentification > &	peptides
	)

static

Removes protein hits from proteins that are not referenced by a peptide in peptides.

◆ updateHitRanks()

static void updateHitRanks ( std::vector< IdentificationType > & ids )

inlinestatic

Updates the hit ranks on all peptide or protein IDs.

◆ updateProteinGroups()

static bool updateProteinGroups	(	std::vector< ProteinIdentification::ProteinGroup > &	groups,
		const std::vector< ProteinHit > &	hits
	)

static

Update protein groups after protein hits were filtered.

Parameters

groups	Input/output protein groups
hits	Available protein hits (all others are removed from the groups)

Returns: Returns whether the groups are still valid (which is the case if only whole groups, if any, were removed).

◆ updateProteinReferences()

static void updateProteinReferences	(	std::vector< PeptideIdentification > &	peptides,
		const std::vector< ProteinIdentification > &	proteins,
		bool	remove_peptides_without_reference = `false`
	)

static

Removes references to missing proteins.

Only PeptideEvidence entries that reference protein hits in proteins are kept in the peptide hits.

If remove_peptides_without_reference is set, peptide hits without any remaining protein reference are removed.

Classes

Public Member Functions

Static Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ IDFilter()

◆ ~IDFilter()

Member Function Documentation

◆ countHits()

◆ extractPeptideSequences()

◆ filterHitsByRank()

◆ filterHitsByScore() [1/2]

◆ filterHitsByScore() [2/2]

◆ filterHitsBySignificance() [1/2]

◆ filterHitsBySignificance() [2/2]

◆ FilterPeptideEvidences()

◆ filterPeptideEvidences()

◆ filterPeptidesByCharge()

◆ filterPeptidesByLength()

◆ filterPeptidesByMZ()

◆ filterPeptidesByMZError()

◆ filterPeptidesByRT()

◆ filterPeptidesByRTPredictPValue()

◆ getBestHit()

◆ keepBestPeptideHits()

◆ keepHitsMatchingProteins() [1/2]

◆ keepHitsMatchingProteins() [2/2]

◆ keepMatchingItems()

◆ keepNBestHits() [1/2]

◆ keepNBestHits() [2/2]

◆ keepPeptidesWithMatchingModifications()

◆ keepPeptidesWithMatchingSequences()

◆ keepUniquePeptidesPerProtein()

◆ removeDecoyHits()

◆ removeDuplicatePeptideHits()

◆ removeEmptyIdentifications()

◆ removeHitsMatchingProteins()

◆ removeMatchingItems()

◆ removePeptidesWithMatchingModifications()

◆ removePeptidesWithMatchingSequences()

◆ removeUnreferencedProteins()

◆ updateHitRanks()

◆ updateProteinGroups()

◆ updateProteinReferences()