An implementation of the X!Tandem HyperScore PSM scoring function. More...

#include <OpenMS/ANALYSIS/ID/HyperScore.h>

Classes
struct	PSMDetail
	compute the (ln transformed) X!Tandem HyperScore overload that returns some additional information on the match More...

Public Types
typedef std::pair< Size, double >	IndexScorePair

Static Public Member Functions
static double	compute (double fragment_mass_tolerance, bool fragment_mass_tolerance_unit_ppm, const PeakSpectrum &exp_spectrum, const PeakSpectrum &theo_spectrum)
	compute the (ln transformed) X!Tandem HyperScore

static double	computeWithDetail (double fragment_mass_tolerance, bool fragment_mass_tolerance_unit_ppm, const PeakSpectrum &exp_spectrum, const PeakSpectrum &theo_spectrum, PSMDetail &d)

static double	computeMassAccuracy (double fragment_mass_tolerance, bool fragment_mass_tolerance_unit_ppm, const PeakSpectrum &exp_spectrum, const PeakSpectrum &theo_spectrum, double mass_error_sd_ppm, PSMDetail &detail)
	HyperScore with mass-accuracy-weighted fragment evidence.

static double	computeWithDetail (double fragment_mass_tolerance, bool fragment_mass_tolerance_unit_ppm, const PeakSpectrum &exp_spectrum, const PeakSpectrum &theo_spectrum, const std::vector< uint8_t > &theo_ion_series, PSMDetail &d)
	computeWithDetail() and computeMassAccuracy() for a theoretical spectrum without ion names

static double	computeMassAccuracy (double fragment_mass_tolerance, bool fragment_mass_tolerance_unit_ppm, const PeakSpectrum &exp_spectrum, const PeakSpectrum &theo_spectrum, const std::vector< uint8_t > &theo_ion_series, double mass_error_sd_ppm, PSMDetail &detail)
	computeWithDetail() and computeMassAccuracy() for a theoretical spectrum without ion names

static double	compute (double fragment_mass_tolerance, bool fragment_mass_tolerance_unit_ppm, const PeakSpectrum &exp_spectrum, const DataArrays::IntegerDataArray &exp_charges, const PeakSpectrum &theo_spectrum, const DataArrays::IntegerDataArray &theo_charges)

static double	compute (double fragment_mass_tolerance, bool fragment_mass_tolerance_unit_ppm, const PeakSpectrum &exp_spectrum, const DataArrays::IntegerDataArray &exp_charges, const PeakSpectrum &theo_spectrum, const DataArrays::IntegerDataArray &theo_charges, std::vector< double > &intensity_sum)

Static Private Member Functions
static double	logfactorial_ (const int x, int base=2)
	helper to compute the log factorial

Detailed Description

An implementation of the X!Tandem HyperScore PSM scoring function.

Class Documentation

◆ OpenMS::HyperScore::PSMDetail

struct OpenMS::HyperScore::PSMDetail

compute the (ln transformed) X!Tandem HyperScore overload that returns some additional information on the match

Class Members
size_t	matched_prefix_ions = 0	N-terminal ions (a, b, c)
size_t	matched_suffix_ions = 0	C-terminal ions (x, y, z)
double	mean_error = 0.0

Member Typedef Documentation

◆ IndexScorePair

typedef std::pair<Size, double> IndexScorePair

Member Function Documentation

◆ compute() [1/3]

static double compute	(	double	fragment_mass_tolerance,
		bool	fragment_mass_tolerance_unit_ppm,
		const PeakSpectrum &	exp_spectrum,
		const DataArrays::IntegerDataArray &	exp_charges,
		const PeakSpectrum &	theo_spectrum,
		const DataArrays::IntegerDataArray &	theo_charges
	)

static

◆ compute() [2/3]

static double compute	(	double	fragment_mass_tolerance,
		bool	fragment_mass_tolerance_unit_ppm,
		const PeakSpectrum &	exp_spectrum,
		const DataArrays::IntegerDataArray &	exp_charges,
		const PeakSpectrum &	theo_spectrum,
		const DataArrays::IntegerDataArray &	theo_charges,
		std::vector< double > &	intensity_sum
	)

static

◆ compute() [3/3]

static double compute	(	double	fragment_mass_tolerance,
		bool	fragment_mass_tolerance_unit_ppm,
		const PeakSpectrum &	exp_spectrum,
		const PeakSpectrum &	theo_spectrum
	)

static

compute the (ln transformed) X!Tandem HyperScore

the dot product of peak intensities between matching peaks in experimental and theoretical spectrum is calculated

the HyperScore is calculated from the dot product by multiplying by factorials of matching b- and y-ions

Note: Peak intensities of the theoretical spectrum are typically 1 or TIC normalized, but can also be e.g. ion probabilities

Parameters

[in]	fragment_mass_tolerance	mass tolerance applied left and right of the theoretical spectrum peak position
[in]	fragment_mass_tolerance_unit_ppm	Unit of the mass tolerance is: Thomson if false, ppm if true
[in]	exp_spectrum	measured spectrum
[in]	theo_spectrum	theoretical spectrum Peaks need to contain an ion annotation as provided by TheoreticalSpectrumGenerator.

◆ computeMassAccuracy() [1/2]

static double computeMassAccuracy	(	double	fragment_mass_tolerance,
		bool	fragment_mass_tolerance_unit_ppm,
		const PeakSpectrum &	exp_spectrum,
		const PeakSpectrum &	theo_spectrum,
		const std::vector< uint8_t > &	theo_ion_series,
		double	mass_error_sd_ppm,
		PSMDetail &	detail
	)

static

computeWithDetail() and computeMassAccuracy() for a theoretical spectrum without ion names

theo_ion_series holds per peak of theo_spectrum 0 for an N-terminal (a, b, c) and 1 for a C-terminal (x, y, z) ion, as TheoreticalSpectrumGenerator::getIonSeriesSpectrum() provides it; any other value counts as neither. The scores are those of the overloads above for the ion names of the same ions.

Exceptions

Exception::InvalidSize if theo_ion_series does not hold one entry per peak

◆ computeMassAccuracy() [2/2]

static double computeMassAccuracy	(	double	fragment_mass_tolerance,
		bool	fragment_mass_tolerance_unit_ppm,
		const PeakSpectrum &	exp_spectrum,
		const PeakSpectrum &	theo_spectrum,
		double	mass_error_sd_ppm,
		PSMDetail &	detail
	)

static

HyperScore with mass-accuracy-weighted fragment evidence.

A matched ion with signed mass error e (ppm) contributes exp(-0.5*(e/sigma)^2) to its terminal ion count and intensity product. Fractional factorial rewards use max(0, lgamma(count+1)), agreeing with HyperScore for exact matches. The kernel is centred at zero and its width is fixed; it is a ranking statistic, not a PSM p-value.

Parameters

[in]	fragment_mass_tolerance	Positive finite match tolerance (Da or ppm).
[in]	fragment_mass_tolerance_unit_ppm	Whether matching tolerance is in ppm.
[in]	exp_spectrum	Experimental spectrum, sorted by m/z.
[in]	theo_spectrum	Theoretical spectrum, sorted by m/z, with ion names.
[in]	mass_error_sd_ppm	Positive finite Gaussian standard deviation in ppm.
[out]	detail	Unweighted match counts and mean absolute error in matching units.

Returns: Nonnegative weighted log HyperScore; zero without matches.

Exceptions

Exception::InvalidParameter	for invalid tolerance or standard deviation.
Exception::InvalidValue	for incomplete theoretical ion annotations.

◆ computeWithDetail() [1/2]

static double computeWithDetail	(	double	fragment_mass_tolerance,
		bool	fragment_mass_tolerance_unit_ppm,
		const PeakSpectrum &	exp_spectrum,
		const PeakSpectrum &	theo_spectrum,
		const std::vector< uint8_t > &	theo_ion_series,
		PSMDetail &	d
	)

static

computeWithDetail() and computeMassAccuracy() for a theoretical spectrum without ion names

theo_ion_series holds per peak of theo_spectrum 0 for an N-terminal (a, b, c) and 1 for a C-terminal (x, y, z) ion, as TheoreticalSpectrumGenerator::getIonSeriesSpectrum() provides it; any other value counts as neither. The scores are those of the overloads above for the ion names of the same ions.

Exceptions

Exception::InvalidSize if theo_ion_series does not hold one entry per peak

◆ computeWithDetail() [2/2]

static double computeWithDetail	(	double	fragment_mass_tolerance,
		bool	fragment_mass_tolerance_unit_ppm,
		const PeakSpectrum &	exp_spectrum,
		const PeakSpectrum &	theo_spectrum,
		PSMDetail &	d
	)

static

◆ logfactorial_()

static double logfactorial_	(	const int	x,
		int	base = `2`
	)

staticprivate

helper to compute the log factorial

Classes

Public Types

Static Public Member Functions

Static Private Member Functions

Detailed Description

Class Documentation

◆ OpenMS::HyperScore::PSMDetail

Member Typedef Documentation

◆ IndexScorePair

Member Function Documentation

◆ compute() [1/3]

◆ compute() [2/3]

◆ compute() [3/3]

◆ computeMassAccuracy() [1/2]

◆ computeMassAccuracy() [2/2]

◆ computeWithDetail() [1/2]

◆ computeWithDetail() [2/2]

◆ logfactorial_()