`esmraldi.spectraprocessing`¶

Module for the preprocessing of spectra specifically designed for MALDI images

Peak picking

Local realignment procedures

Deisotoping

Module Contents¶

Functions¶

`spectra_sum`(spectra)	Computes the spectrum from the sum of all spectra.
`spectra_mean`(spectra)	Computes the average spectrum.
`spectra_mean_centroided`(spectra[, mzs])
`spectra_max`(spectra)	Computes the maximum intensity for each abscissa.
`spectra_min`(spectra)	Computes the minimum intensity for each abscissa.
`spectra_peak_indices`(spectra[, prominence, wlen])	Estimates and extracts significant peaks in the spectra
`spectra_peak_indices_adaptative`(spectra[, factor, wlen])	Estimates and extracts significant peaks in the spectra
`spectra_peak_mzs_adaptative`(spectra[, factor, wlen])	Estimates and extracts significant peaks in the spectra
`spectra_peak_indices_adaptative_noiselevel`(spectra[, ...])	Estimates and extracts significant peaks in the spectra
`spectra_peak_mzs_adaptative_noiselevel`(spectra[, ...])	Estimates and extracts significant peaks in the spectra
`peak_indices`(data, prominence, wlen[, distance])	Estimates and extracts significant peaks
`peak_indices_cwt`(data, factor, widths)	Peak indices using continuous wavelet
`spectra_peak_mzs_cwt`(spectra, factor, widths)	Peak detection using the Continuous Wavelet Transform.
`same_mz_axis`(spectra[, tol])	Generates spectra with common m/z values.
`tic_values`(spectra)
`normalization_tic`(spectra[, inplace])	TIC (total ion count) normalization.
`normalization_sic`(spectra, indices_peaks[, width_peak])	SIC (selective ion count) normalization.
`index_groups`(indices[, step, is_ppm])	Makes groups of indices.
`index_groups_start_end`(indices[, step, is_ppm])
`peak_reference_indices_group`(group)	Extracts the reference peak in a group,
`peak_reference_indices_groups`(groups)	Extracts the reference peaks for several groups.
`peak_reference_indices_median`(groups)	Extracts the reference peak in a group
`width_peak_mzs`(aligned_mzs, groups[, default])	Computes the width of a peak
`width_peak_indices`(indices, full_indices)	Computes the width of a peak
`closest_peak`(num, indices_to_width)	Extracts the closest peak of index num.
`min_step`(mzs, max_len[, starting_step, incr_step])
`realign_reducing`(out_spectra, spectra[, step, is_ppm])
`realign_mean_spectrum`(mzs, intensities, all_mzs[, ...])
`realign_tree`(spectra, mzs, mean_spectra[, step, is_ppm])
`realign_wrt_peaks_mzs_generic`(spectra, aligned_mzs)
`realign_wrt_peaks_mzs`(spectra, aligned_mzs, full_mzs, ...)	Realign spectra to reference peaks.
`realign_wrt_peaks`(spectra, aligned_peaks, full_peaks, ...)	Realign spectra to reference peaks
`realign_indices`(spectra, indices[, reference, ...])	Alignment function.
`realign_mzs`(spectra, mzs[, reference, nb_occurrence, ...])	Alignment function.
`realign_generic`(spectra, peaks[, step, is_ppm])
`neighbours`(index, n, spectra)	Right-sided neighbours of a point in a spectrum.
`forward_derivatives`(peaks)	Forward derivatives from peak value distribution.
`find_isotopic_pattern`(neighbours, tolerance, nb_charges)	Extracts isotopic pattern based on mz similarity
`find_isotopic_pattern_theoretical_difference`(...[, is_ppm])
`peaks_max_intensity_isotopic_pattern`(pattern)	Finds the peak with maximum intensity in
`peaks_derivative_isotopic_pattern`(pattern)	Finds the peak where the sign of the derivative changes
`isotopes_from_pattern`(pattern, peaks_in_pattern)	Find isotopes from a pattern.
`mz_second_isotope_most_abundant`(average_distribution)	Determines where the second isotope becomes the most abundant
`peak_to_index`(peak, pattern)	Gets the index of a peak in a pattern
`deisotoping_simple`(spectra[, tolerance, ...])	Simple deisotoping depending on the mass of the
`deisotoping_simple_reference`(spectra[, th_diff, ...])	Simple deisotoping depending on the mass of the
`deisotoping_reference_indices`(peaks[, th_diff, ...])
`subtract_spectra`(target, source)
`extract_mean_spectra_coordinates`(spectra, coordinates, ...)

esmraldi.spectraprocessing.spectra_sum(spectra)¶

Computes the spectrum from the sum of all spectra.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array

Returns

np.ndarray: Sum of spectra

esmraldi.spectraprocessing.spectra_mean(spectra)¶

Computes the average spectrum.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array

Returns

np.ndarray: Mean spectrum

esmraldi.spectraprocessing.spectra_mean_centroided(spectra, mzs=None)¶

esmraldi.spectraprocessing.spectra_max(spectra)¶

Computes the maximum intensity for each abscissa.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array

Returns

np.ndarray: Max spectrum

esmraldi.spectraprocessing.spectra_min(spectra)¶

Computes the minimum intensity for each abscissa.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array

Returns

np.ndarray: Max spectrum

esmraldi.spectraprocessing.spectra_peak_indices(spectra, prominence=50, wlen=10)¶

Estimates and extracts significant peaks in the spectra by using the prominence (height of the peak relative to the nearest higher peak).

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
prominence: int: threshold on prominence

Returns

np.ndarray: Peak indices relative to spectra

esmraldi.spectraprocessing.spectra_peak_indices_adaptative(spectra, factor=1, wlen=10)¶

Estimates and extracts significant peaks in the spectra by using the local prominence (height of the peak relative to the background noise).

Background noise is estimated as the standard deviation of the signal over a window of size wlen.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
factor: float: prominence factor
wlen: int: size of the window

Returns

np.ndarray: Peak indices relative to spectra

esmraldi.spectraprocessing.spectra_peak_mzs_adaptative(spectra, factor=1, wlen=10)¶

Estimates and extracts significant peaks in the spectra by using the local prominence (height of the peak relative to the background noise).

Background noise is estimated as the standard deviation of the signal over a window of size wlen.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
factor: float: prominence factor
wlen: int: size of the window

Returns

np.ndarray: Peaks m/z

esmraldi.spectraprocessing.spectra_peak_indices_adaptative_noiselevel(spectra, factor=1, noise_level=1, wlen=10)¶

Estimates and extracts significant peaks in the spectra with specified noise level(s), by using the local prominence (height of the peak relative to the background noise)

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
factor: float: prominence factor
noise_level: float or list: noise level
wlen: int: size of the window

Returns

np.ndarray: Peaks m/z

esmraldi.spectraprocessing.spectra_peak_mzs_adaptative_noiselevel(spectra, factor=1, noise_level=1, wlen=10)¶

Estimates and extracts significant peaks in the spectra with specified noise level(s), by using the local prominence (height of the peak relative to the background noise)

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
factor: float: prominence factor
noise_level: float or list: noise level
wlen: int: size of the window

Returns

np.ndarray: Peaks m/z

esmraldi.spectraprocessing.peak_indices(data, prominence, wlen, distance=1)¶

Estimates and extracts significant peaks in the spectrum, by using the prominence (height of the peak relative to the nearest higher peak).

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
prominence: int: threshold on prominence

Returns

np.ndarray: Peak indices relative to spectrum

esmraldi.spectraprocessing.peak_indices_cwt(data, factor, widths)¶

Peak indices using continuous wavelet transform.

Parameters

data: np.ndarray: Spectra as [mz*I] array
factor: float: Threshold SNR
widths: list: scales

Returns

np.ndarray: Peak indices relative to spectrum

esmraldi.spectraprocessing.spectra_peak_mzs_cwt(spectra, factor, widths)¶

Peak detection using the Continuous Wavelet Transform.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
factor: float: CWT threshold
widths: list: Wavelet widths

Returns

np.ndarray: Detected peak m/z ratios

esmraldi.spectraprocessing.same_mz_axis(spectra, tol=0)¶

Generates spectra with common m/z values.

Missing intensity values are added as np.nan.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
tol: float: Tolerance to consider when two species are the same

Returns

np.ndarray: Spectra as [mz*I] array

esmraldi.spectraprocessing.tic_values(spectra)¶

esmraldi.spectraprocessing.normalization_tic(spectra, inplace=False)¶

TIC (total ion count) normalization.

Divides each intensity in a spectrum by the sum of all its intensities.

Parameters

spectra: np.ndarray: spectra as [mz*I] array

Returns

np.ndarray: normalized spectrum

esmraldi.spectraprocessing.normalization_sic(spectra, indices_peaks, width_peak=10)¶

SIC (selective ion count) normalization.

Defined as : TIC - sum of peaks of high intensities Peaks are given with indices_peaks.

Parameters

spectra: np.ndarray: spectra as [mz*I] array
indices_peaks: np.ndarray: indices peaks
width_peak: int: average width of peaks

Returns

np.ndarray: normalized spectrum

esmraldi.spectraprocessing.index_groups(indices, step=1, is_ppm=False)¶

Makes groups of indices.

For realignment and spatial selection.

Parameters

indices: list: list of peak indices
step: int: threshold in indices to create groups

Returns

list: groups=list of list of peak indices

esmraldi.spectraprocessing.index_groups_start_end(indices, step=1, is_ppm=False)¶

esmraldi.spectraprocessing.peak_reference_indices_group(group)¶

Extracts the reference peak in a group, i.e. the most frequent in a group.

Parameters

group: list: list of peak indices

esmraldi.spectraprocessing.peak_reference_indices_groups(groups)¶

Extracts the reference peaks for several groups.

Parameters

groups: list: groups=list of list of peak indices

Returns

list: list of reference peak indices

esmraldi.spectraprocessing.peak_reference_indices_median(groups)¶

Extracts the reference peak in a group as the median peak.

Parameters

groups: list: list of peak indices

Returns

list: list of reference peak indices

esmraldi.spectraprocessing.width_peak_mzs(aligned_mzs, groups, default=0.001)¶

Computes the width of a peak by computing the difference in m/z between the upper and lower bounds in the group.

Parameters

aligned_mzs: list: list of m/z
groups: list: list of peak indices
default: float: default width of the peak

Returns

dict: maps m/z to widths

esmraldi.spectraprocessing.width_peak_indices(indices, full_indices)¶

Computes the width of a peak by checking neighboring indices.

Parameters

indices: np.ndarray: peak indices on the histogram of peaks
full_indices: np.ndarray: peak indices on the spectra

Returns

dict:: peak indices to corresponding width

esmraldi.spectraprocessing.closest_peak(num, indices_to_width)¶

Extracts the closest peak of index num.

Parameters

num: int: index
indices_to_width: dict: dictionary mapping indices to width

Returns

tuple: mz and associated width

esmraldi.spectraprocessing.min_step(mzs, max_len, starting_step=0.0005, incr_step=0.0005)¶

esmraldi.spectraprocessing.realign_reducing(out_spectra, spectra, step=0.0005, is_ppm=False)¶

esmraldi.spectraprocessing.realign_mean_spectrum(mzs, intensities, all_mzs, step=0.0005, is_ppm=False, return_stats=False)¶

esmraldi.spectraprocessing.realign_tree(spectra, mzs, mean_spectra, step=0.0005, is_ppm=False)¶

esmraldi.spectraprocessing.realign_wrt_peaks_mzs_generic(spectra, aligned_mzs)¶

esmraldi.spectraprocessing.realign_wrt_peaks_mzs(spectra, aligned_mzs, full_mzs, indices_to_width)¶

Realign spectra to reference peaks.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
aligned_mzs: list: reference mz peaklist
full_mzs: np.ndarray: complete mz peaklist over all spectra
indices_to_width: dict: mz to width

Returns

list: realigned spectra

esmraldi.spectraprocessing.realign_wrt_peaks(spectra, aligned_peaks, full_peaks, indices_to_width)¶

Realign spectra to reference peaks from indices.

Parameters

spectra: np.ndarray: Spectra as [mz*I] array
aligned_mzs: list: reference mz peaklist
full_mzs: np.ndarray: complete mz peaklist over all spectra
indices_to_width: dict: indices to width

Returns

list: realigned spectra indices

esmraldi.spectraprocessing.realign_indices(spectra, indices, reference='frequence', nb_occurrence=4, step=0.02, is_ppm=False)¶

Alignment function.

First extracts the peaks on all spectra, then extracts the reference peaks and maps each peak to its closest reference peak.

Parameters

spectra: np.ndarray: spectra
indices: np.ndarray: indices of peaks relative to spectra

Returns

list: realigned spectra

esmraldi.spectraprocessing.realign_mzs(spectra, mzs, reference='frequence', nb_occurrence=4, step=0.02, is_ppm=False)¶

Alignment function.

First extracts the peaks on all spectra based on local prominence, then extracts the reference peaks and maps each peak to its closest reference peak.

Parameters

spectra: np.ndarray: spectra
mzs: np.ndarray: peaklist of mzs ratio

Returns

list: realigned spectra

esmraldi.spectraprocessing.realign_generic(spectra, peaks, step=np.inf, is_ppm=False)¶

esmraldi.spectraprocessing.neighbours(index, n, spectra)¶

Right-sided neighbours of a point in a spectrum.

Parameters

index: int: index to search neighbours from
n: int: number of neighbours
spectra: np.ndarray: spectrum

Returns

np.ndarray: neighbours

esmraldi.spectraprocessing.forward_derivatives(peaks)¶

Forward derivatives from peak value distribution.

Parameters

peaks: list: peaklist

Returns

list: derivatives

esmraldi.spectraprocessing.find_isotopic_pattern(neighbours, tolerance, nb_charges)¶

Extracts isotopic pattern based on mz similarity and max number of charges.

Parameters

neighbours: np.ndarray: neighbouring mz in spectra
tolerance: float: acceptable mz delta for a peak to be considered isotopic
nb_charges: int: maximum number of charges

Returns

list: pattern peaklist

esmraldi.spectraprocessing.find_isotopic_pattern_theoretical_difference(neighbours, th_diff, tolerance, nb_charges, is_ppm=True)¶

esmraldi.spectraprocessing.peaks_max_intensity_isotopic_pattern(pattern)¶

Finds the peak with maximum intensity in the isotopic pattern.

Parameters

pattern: list: pattern peaklist

Returns

list: peak with maximum intensity

esmraldi.spectraprocessing.peaks_derivative_isotopic_pattern(pattern)¶

Finds the peak where the sign of the derivative changes from negative to positive from the pattern intensities.

Parameters

pattern: list: pattern peaklist

Returns

list: peaks where the derivative sign changes

esmraldi.spectraprocessing.isotopes_from_pattern(pattern, peaks_in_pattern)¶

Find isotopes from a pattern.

Parameters

pattern: list: pattern peaklist
peaks_in_pattern: list: peaks that correspond to other species in the pattern

Returns

np.ndarray: isotopes in the pattern

esmraldi.spectraprocessing.mz_second_isotope_most_abundant(average_distribution)¶

Determines where the second isotope becomes the most abundant from a given distribution.

Parameters

average_distribution: dict: maps atom mass to its average abundance

Returns

float: mass at which the second isotope is the most abundant

esmraldi.spectraprocessing.peak_to_index(peak, pattern)¶

Gets the index of a peak in a pattern

Parameters

peak: float: mass ratio
pattern: list: pattern peaklist

Returns

int: index of peak

esmraldi.spectraprocessing.deisotoping_simple(spectra, tolerance=0.1, nb_neighbours=8, nb_charges=5, average_distribution={})¶

Simple deisotoping depending on the mass of the secondmost abundant isotope:

Before this mass: uses the peak with max intensity as reference

After this mass: use the peak where the sign of the derivative changes

Parameters

spectra: np.ndarray: peaklist
tolerance: float: acceptable mz delta
nb_neighbours: int: size of patterns
nb_charges: int: maximum number of charges in isotopic pattern
average_distribution: dict: maps atom mass to its average abundance

Returns

np.ndarray: deisotoped spectra

esmraldi.spectraprocessing.deisotoping_simple_reference(spectra, th_diff=1.00335, tolerance=14, nb_neighbours=8, nb_charges=5, is_ppm=True)¶

Simple deisotoping depending on the mass of the secondmost abundant isotope:

Before this mass: uses the peak with max intensity as reference

After this mass: use the peak where the sign of the derivative changes

Parameters

spectra: np.ndarray: peaklist
tolerance: float: acceptable mz delta
nb_neighbours: int: size of patterns
nb_charges: int: maximum number of charges in isotopic pattern
average_distribution: dict: maps atom mass to its average abundance

Returns

np.ndarray: deisotoped spectra

esmraldi.spectraprocessing.deisotoping_reference_indices(peaks, th_diff=1.00335, tolerance=14, nb_neighbours=8, nb_charges=5, is_ppm=True)¶

esmraldi.spectraprocessing.subtract_spectra(target, source)¶

esmraldi.spectraprocessing.extract_mean_spectra_coordinates(spectra, coordinates, mzs, is_subtract, mean_spectra_matrix)¶

esmraldi.spectraprocessing¶

Module Contents¶

Functions¶

`esmraldi.spectraprocessing`¶