FLASHHelperClasses::PrecalculatedAveragine Class Reference

Pre-binned averagine isotope-distribution cache used by FLASHDeconv for fast cosine-similarity scoring. More...

#include <OpenMS/ANALYSIS/TOPDOWN/FLASHHelperClasses.h>

Collaboration diagram for FLASHHelperClasses::PrecalculatedAveragine:

Public Member Functions
	PrecalculatedAveragine ()=default
	Default-construct an empty cache (no bins populated; the accessors are not safe until the parameterised ctor is used).
	PrecalculatedAveragine (double min_mass, double max_mass, double delta, CoarseIsotopePatternGenerator &generator, bool use_RNA_averagine, double decoy_iso_distance=-1)
	Precompute averagine isotope distributions across min_mass .. max_mass in steps of delta.
	PrecalculatedAveragine (const PrecalculatedAveragine &)=default
	Copy constructor.
	PrecalculatedAveragine (PrecalculatedAveragine &&other) noexcept=default
	Move constructor.
PrecalculatedAveragine &	operator= (const PrecalculatedAveragine &pc)=default
	Copy assignment.
PrecalculatedAveragine &	operator= (PrecalculatedAveragine &&pc) noexcept=default
	Move assignment.
	~PrecalculatedAveragine ()=default
	Destructor.
IsotopeDistribution	get (double mass) const
	Return the cached trimmed + L2-normalised isotope distribution for the bin containing mass.
size_t	getMaxIsotopeIndex () const
	Return the externally-set isotope-index cap.
void	setMaxIsotopeIndex (int index)
	Set the isotope-index cap consulted by external callers. Does not influence the cached distributions or other accessors.
Size	getLeftCountFromApex (double mass) const
	Return the number of significant isotopes to the left of the apex for the bin containing mass.
Size	getRightCountFromApex (double mass) const
	Return the number of significant isotopes to the right of the apex for the bin containing mass.
Size	getApexIndex (double mass) const
	Return the index of the most-abundant isotope inside the trimmed distribution for the bin containing mass.
Size	getLastIndex (double mass) const
	Return apex_index + right_count_from_apex for the bin containing mass — the index of the rightmost significant isotope.
double	getAverageMassDelta (double mass) const
	Return (average_mass - monoisotopic_mass) for the bin's trimmed distribution.
double	getMostAbundantMassDelta (double mass) const
	Return (most_abundant_mass - monoisotopic_mass) for the bin's trimmed distribution.
double	getSNRMultiplicationFactor (double mass) const
	Return the SNR multiplication factor for the bin containing mass.

Private Member Functions
Size	massToIndex_ (double mass) const
	Convert mass to a bin index. Clamps to [0, isotopes_.size()-1] for out-of-range inputs (no throw).

Private Attributes
std::vector< IsotopeDistribution >	isotopes_
	Per-bin isotope distribution. After trimming and L2 normalisation in the constructor.
std::vector< double >	norms_
	Per-bin L2 norm — populated and consumed alongside isotopes_ for fast cosine scoring (vestigial: not consumed by the public API in this header).
std::vector< double >	average_mono_mass_difference_
	Per-bin difference between the average mass and the monoisotopic mass of the trimmed distribution.
std::vector< double >	abundant_mono_mass_difference_
	Per-bin difference between the most-abundant-isotope mass and the monoisotopic mass.
std::vector< double >	snr_mul_factor_
	Per-bin SNR multiplication factor: (sum-of-normalised-intensities)^2 of the trimmed distribution.
std::vector< int >	left_count_from_apex_
	Per-bin count of significant isotopes on the left side of the apex (running maximum across bins; see the PrecalculatedAveragine parameterised constructor for the unusual running-max semantics).
std::vector< int >	right_count_from_apex_
	Per-bin count of significant isotopes on the right side of the apex (running maximum across bins; same caveat as left_count_from_apex_).
std::vector< Size >	apex_index_
	Per-bin index of the most-abundant isotope inside the trimmed distribution.
Size	max_isotope_index_
	Cap on the isotope index used externally by FLASHDeconv. NOT initialised by the constructor — call setMaxIsotopeIndex before reading via getMaxIsotopeIndex.
double	mass_interval_
	Bin step size (the delta passed to the constructor).
double	min_mass_
	Lower bound of the cached mass range (the min_mass passed to the constructor).

Detailed Description

Pre-binned averagine isotope-distribution cache used by FLASHDeconv for fast cosine-similarity scoring.

Builds a table keyed by a mass bin index — one IsotopeDistribution plus its derived quantities (apex index, left/right counts of significant isotopes, average / most-abundant mass offsets, an SNR multiplication factor) per mass bin from min_mass to max_mass at step delta. The constructor performs all the per-mass work up front; the accessors are O(1) table lookups via massToIndex_(mass) (rounded to the nearest bin).

Bin lookup silently clamps: any mass below min_mass maps to bin 0 and any mass above max_mass maps to the last bin, so the accessors never throw on out-of-range input.

Constructor & Destructor Documentation

PrecalculatedAveragine() [1/4]

PrecalculatedAveragine ( )

default

Default-construct an empty cache (no bins populated; the accessors are not safe until the parameterised ctor is used).

PrecalculatedAveragine() [2/4]

PrecalculatedAveragine	(	double	min_mass,
		double	max_mass,
		double	delta,
		CoarseIsotopePatternGenerator &	generator,
		bool	use_RNA_averagine,
		double	decoy_iso_distance = -1
	)

Precompute averagine isotope distributions across min_mass .. max_mass in steps of delta.

For each bin mass m = i*delta in the range, calls CoarseIsotopePatternGenerator::estimateFromPeptideMonoWeight(m) (or estimateFromRNAMonoWeight(m) when use_RNA_averagine is true) and applies the following processing:

• Trimming: iteratively drops the least-intense end peak (left or right) until the trimmed-out power is below 0.0001 of the total, with a minimum kept length of 2 peaks. Trimmed peaks have their intensity zeroed out, not removed from the distribution.
• L2 normalisation: divides every remaining peak's intensity by sqrt(total_pwr) of the kept set.
• Decoy mode: when decoy_iso_distance is > 0, every peak's m/z is multiplied by decoy_iso_distance before trimming, then the distribution is re-sorted by mass and renormalised. This produces a "stretched" pattern whose inter-isotope spacing is decoy_iso_distance times the natural spacing.
• Left/right counts: clamped to a floor of 2 each, and then accumulated as a running maximum across bins — so all bins end up with the same width (the maximum seen so far during construction).

Per bin, the constructor records the trimmed distribution, the apex index, the running-max left/right counts, the average-mono and most-abundant-mono mass deltas, and the SNR multiplication factor (sum-of-normalised-intensities)^2.

Parameters

[in]	min_mass	Lower bound of the cached mass range; bins are built starting at the first multiple of delta that is >= min_mass.
[in]	max_mass	Upper bound of the cached mass range; bin generation stops at the first multiple of delta that exceeds max_mass.
[in]	delta	Bin step size.
[in,out]	generator	Isotope-pattern generator used to compute each per-mass distribution.
[in]	use_RNA_averagine	If true, use the RNA-mass averagine; otherwise use peptide averagine.
[in]	decoy_iso_distance	> 0 enables decoy mode (per-peak m/z multiplied by this value); <= 0 disables it. Default -1.

PrecalculatedAveragine() [3/4]

PrecalculatedAveragine ( const PrecalculatedAveragine &  )

default

Copy constructor.

PrecalculatedAveragine() [4/4]

PrecalculatedAveragine ( PrecalculatedAveragine &&  other )

defaultnoexcept

Move constructor.

~PrecalculatedAveragine()

~PrecalculatedAveragine ( )

default

Destructor.

Member Function Documentation

get()

IsotopeDistribution get

(

double

mass

)

const

Return the cached trimmed + L2-normalised isotope distribution for the bin containing mass.

Parameters

[in]

mass

Input mass to query. Out-of-range values are silently clamped to [min_mass, max_mass].

Returns

Cached isotope distribution for the resolved bin.

getApexIndex()

Size getApexIndex

(

double

mass

)

const

Return the index of the most-abundant isotope inside the trimmed distribution for the bin containing mass.

Parameters

[in]

mass

Input mass to query. Out-of-range values are silently clamped to [min_mass, max_mass].

Returns

Position of the apex peak inside the trimmed distribution.

getAverageMassDelta()

double getAverageMassDelta

(

double

mass

)

const

Return (average_mass - monoisotopic_mass) for the bin's trimmed distribution.

Parameters

[in]

mass

Input mass to query. Out-of-range values are silently clamped to [min_mass, max_mass].

Returns

Mass difference between the trimmed distribution's average mass and its monoisotopic mass.

getLastIndex()

Size getLastIndex

(

double

mass

)

const

Return apex_index + right_count_from_apex for the bin containing mass — the index of the rightmost significant isotope.

Parameters

[in]

mass

Input mass to query. Out-of-range values are silently clamped to [min_mass, max_mass].

Returns

Index of the rightmost significant isotope.

getLeftCountFromApex()

Size getLeftCountFromApex

(

double

mass

)

const

Return the number of significant isotopes to the left of the apex for the bin containing mass.

Note

The stored value is a running maximum accumulated across bins during construction, so it grows monotonically with mass (see the parameterised constructor's left/right-count semantics).

Parameters

[in]

mass

Input mass to query. Out-of-range values are silently clamped to [min_mass, max_mass].

Returns

Running-max left count of significant isotopes.

getMaxIsotopeIndex()

size_t getMaxIsotopeIndex

(

)

const

Return the externally-set isotope-index cap.

Warning

The constructor does not initialise max_isotope_index_; reading this before setMaxIsotopeIndex has been called returns an uninitialised value.

Returns

The most recent value passed to setMaxIsotopeIndex.

getMostAbundantMassDelta()

double getMostAbundantMassDelta

(

double

mass

)

const

Return (most_abundant_mass - monoisotopic_mass) for the bin's trimmed distribution.

Parameters

[in]

mass

Input mass to query. Out-of-range values are silently clamped to [min_mass, max_mass].

Returns

Mass difference between the trimmed distribution's most-abundant-isotope mass and its monoisotopic mass.

getRightCountFromApex()

Size getRightCountFromApex

(

double

mass

)

const

Return the number of significant isotopes to the right of the apex for the bin containing mass.

Note

Same running-maximum caveat as getLeftCountFromApex.

Parameters

[in]

mass

Input mass to query. Out-of-range values are silently clamped to [min_mass, max_mass].

Returns

Running-max right count of significant isotopes.

getSNRMultiplicationFactor()

double getSNRMultiplicationFactor

(

double

mass

)

const

Return the SNR multiplication factor for the bin containing mass.

(sum-of-normalised-intensities)^2 of the trimmed distribution. Used by FLASHDeconv for fast SNR computation.

Parameters

[in]

mass

Input mass to query. Out-of-range values are silently clamped to [min_mass, max_mass].

Returns

Squared sum of normalised peak intensities for the bin.

massToIndex_()

Size massToIndex_ ( double  mass ) const

private

Convert mass to a bin index. Clamps to [0, isotopes_.size()-1] for out-of-range inputs (no throw).

operator=() [1/2]

PrecalculatedAveragine & operator= ( const PrecalculatedAveragine &  pc )

default

Copy assignment.

operator=() [2/2]

PrecalculatedAveragine & operator= ( PrecalculatedAveragine &&  pc )

defaultnoexcept

Move assignment.

setMaxIsotopeIndex()

void setMaxIsotopeIndex

(

int

index

)

Set the isotope-index cap consulted by external callers. Does not influence the cached distributions or other accessors.

Parameters

[in]

index

New isotope-index cap value.

Member Data Documentation

abundant_mono_mass_difference_

std::vector<double> abundant_mono_mass_difference_

private

Per-bin difference between the most-abundant-isotope mass and the monoisotopic mass.

apex_index_

std::vector<Size> apex_index_

private

Per-bin index of the most-abundant isotope inside the trimmed distribution.

average_mono_mass_difference_

std::vector<double> average_mono_mass_difference_

private

Per-bin difference between the average mass and the monoisotopic mass of the trimmed distribution.

isotopes_

std::vector<IsotopeDistribution> isotopes_

private

Per-bin isotope distribution. After trimming and L2 normalisation in the constructor.

left_count_from_apex_

std::vector<int> left_count_from_apex_

private

Per-bin count of significant isotopes on the left side of the apex (running maximum across bins; see the PrecalculatedAveragine parameterised constructor for the unusual running-max semantics).

mass_interval_

double mass_interval_

private

Bin step size (the delta passed to the constructor).

max_isotope_index_

Size max_isotope_index_

private

Cap on the isotope index used externally by FLASHDeconv. NOT initialised by the constructor — call setMaxIsotopeIndex before reading via getMaxIsotopeIndex.

min_mass_

double min_mass_

private

Lower bound of the cached mass range (the min_mass passed to the constructor).

norms_

std::vector<double> norms_

private

Per-bin L2 norm — populated and consumed alongside isotopes_ for fast cosine scoring (vestigial: not consumed by the public API in this header).

right_count_from_apex_

std::vector<int> right_count_from_apex_

private

Per-bin count of significant isotopes on the right side of the apex (running maximum across bins; same caveat as left_count_from_apex_).

snr_mul_factor_

std::vector<double> snr_mul_factor_

private

Per-bin SNR multiplication factor: (sum-of-normalised-intensities)^2 of the trimmed distribution.