Main Content

msnorm

Normalize set of signals with peaks

Description

example

yOut = msnorm(X,Intensities) normalizes a group of signals with peaks by standardizing the area under the curve (AUC) to the group median and returns the normalized data yOut.

example

[yOut,normParams] = msnorm(X,Intensities) also returns the normalization parameters normParams, which you can use to normalize another group of signals.

example

yOut = msnorm(X,Intensities,NormParameters) uses the parameter information NormParameters from a previous normalization to normalize a new set of signals. The function uses the same parameters to select the separation-unit positions and output scale from the previous normalization. If you specified a consensus proportion using the 'Consensus' name-value pair argument in the previous normalization, the function selects no new separation-unit positions and performs normalization using the same separation-unit positions.

example

[___] = msnorm(X,Intensities,Name,Value) uses additional options specified by one or more name-value pair arguments and returns any of the output arguments in previous syntaxes. For example, out = msnorm(X,Y,'Quantile',[0.9 1]) sets the lower (0.9) and upper (1) quantile limit to use only the largest 10% of intensities in each signal to compute the AUC.

Examples

collapse all

This example shows how to normalize the area under the curve of every mass spectrum from the mass spec data.

Load a MAT-file, included with the Bioinformatics Toolbox™ software, that contains sample mass spec data, including MZ_lo_res, a vector of m/z values, and Y_lo_res, a matrix of intensity values.

load sample_lo_res

Create a subset (four signals) of the data.

MZ = MZ_lo_res;
Y = Y_lo_res(:,[1 2 5 6]);

Plot the four spectra.

plot(MZ, Y)
axis([-1000 20000 -20 105])
xlabel('Mass-charge Ratio')
ylabel('Relative Ion Intensities')
title('Original Spectra')

Normalize the area under the curve (AUC) of every spectrum to the median, eliminating low-mass (m/z < 1,000) noise, and post-rescaling such that the maximum intensity is 100. Plot the four spectra.

Y1 = msnorm(MZ,Y,'Limits',[1000 inf],'Max',100);
plot(MZ, Y1)
axis([-1000 20000 -20 105])
xlabel('Mass-charge Ratio')
ylabel('Relative Ion Intensities')
title('AUC Normalized Spectra')

This example shows how to normalize the ion intensity of every spectrum from the mass spec data.

Load a MAT-file, included with the Bioinformatics Toolbox™ software, that contains sample mass spec data, including MZ_lo_res, a vector of m/z values, and Y_lo_res, a matrix of intensity values.

load sample_lo_res

Create a subset (four signals) of the data.

MZ = MZ_lo_res;
Y = Y_lo_res(:,[1 2 5 6]);

Normalize the ion intensity of every spectrum to the maximum intensity of the single highest peak from any of the spectra in the range above 1000 m/z. Plot the four spectra.

Y2 = msnorm(MZ,Y,'QUANTILE', [1 1],'LIMITS',[1000 inf]);
plot(MZ, Y2)
axis([-1000 20000 -20 105])
xlabel('Mass-charge Ratio')
ylabel('Relative Ion Intensities')
title('Maximum-Intensity Normalized Spectra')

This example shows how to perform quantile normalization for mass spec data.

Load a MAT-file, included with the Bioinformatics Toolbox™ software, that contains sample mass spec data, including MZ_lo_res, a vector of m/z values, and Y_lo_res, a matrix of intensity values.

load sample_lo_res

Create a subset (four signals) of the data.

MZ = MZ_lo_res;
Y = Y_lo_res(:,[1 2 5 6]);

Normalize using the data in the m/z regions where the intensities are within the fourth quartile in at least 90% of the spectrograms. Note that you can use the normalization parameters in the second output to normalize another set of data in the same m/z regions. Plot the four spectra.

[Y3,S] = msnorm(MZ,Y,'Quantile',[0.75 1],'Consensus',0.9);
area(MZ,S.Xh.*1000,'LineStyle','None','FaceColor',[.8 .8 .8])
hold on
plot(MZ, Y3)
hold off
axis([-1000 20000 -20 105])
xlabel('Mass-charge Ratio')
ylabel('Relative Ion Intensities')
title('Fourth-quartile Normalized Spectra')

Use the normalization parameters in the second output of the previous step to normalize a different subset of data (four signals) using the data in the same m/z regions as the previous data set. Plot the four spectra.

Y4 = msnorm(MZ,Y_lo_res(:,[3 4 7 8]),S);
 
area(MZ,S.Xh.*1000,'LineStyle','None','FaceColor',[.8 .8 .8])
hold on
plot(MZ, Y4)
hold off
axis([-1000 20000 -20 105])
xlabel('Mass-charge Ratio')
ylabel('Relative Ion Intensities')
title('Fourth-quartile Normalized Spectra')

Input Arguments

collapse all

Vector of separation-unit values for a set of signals with peaks, specified as a vector.

Data Types: double

Intensity values for a set of peaks that share the same separation-unit range, specified as a matrix. Each row is a separation-unit value and each column is either a set of signals with peaks or a retention time. The number of rows in Intensities must equal the number of elements in the input vector X.

Data Types: double

Normalization parameters to normalize another group of signals, specified as a structure. NormParameters is a structure returned by msnorm from a previous normalization call.

Name-Value Arguments

Specify optional pairs of arguments as Name1=Value1,...,NameN=ValueN, where Name is the argument name and Value is the corresponding value. Name-value arguments must appear after other arguments, but the order of the pairs does not matter.

Before R2021a, use commas to separate each name and value, and enclose Name in quotes.

Example: out = msnorm(X,Y,'Quantile',[0.9 1]) sets the lower (0.9) and upper (1) quantile limit to use only the largest 10% of intensities in each signal to compute the AUC.

Quantile limits to reduce the set of separation-unit values in X, specified as a 1-by-2 vector or a scalar between 0 and 1.

If you specify a vector, the first element is the lower limit and the second element is the upper limit. For example, [0.9 1] means that the function uses only the largest 10% of intensities in each signal to compute the AUC. The default value [0 1] means that the function uses the whole AUC, instead of limiting the intensities to a particular quantile.

If you specify a scalar value, it represents the lower quantile limit. The upper quantile limit is automatically set to 1.

Example: 'Quantile',[0.8 1]

Data Types: double

Separation-unit range to pick normalization points, specified as a 1-by-2 vector. The default value [min(X) max(X)] selects all available points from X. If you specify a lower or upper limit as a value that is not within the available range [min(X) max(X)], the function sets the lower limit to min(X) and the upper limit to max(X).

This parameter is useful to eliminate noise from the AUC calculation. For instance, you can exclude the matrix noise that appears in the low-mass region (m/z values less than 1000) of a SELDI mass spectrometer by setting the limit to [1000 max(X)].

Example: 'Limits',[900 max(X)]

Data Types: double

Minimal percentage of intensity values within the quantile limits that a separation-unit position must have to be included in the AUC calculation, specified as a scalar between 0 and 1. The same separation-unit positions are then used to normalize all the signals. Use this parameter to eliminate low-intensity peaks and noise from the normalization.

For instance, to select m/z regions whose intensities are within the third quantile in at least 90% of the spectrograms, set 'Quantile' and 'Consensus' as follows: yOut = msnorm(MZ,Y,'Quantile',[0.5 0.75],'Consensus',0.9).

Example: 'Consensus',0.8

Data Types: double

Method for normalizing the AUC of every signal, specified as 'Median' or 'Mean'.

Example: 'Method','Mean'

Data Types: char | string

Overall maximum intensity to scale to after normalizing each signal individually, specified as a scalar. If you do not specify this parameter, no postscaling is performed.

Note

If you specify this value and also set 'Quantile' to [1 1], then a single point (peak height of the tallest peak) is normalized to the specified maximum value.

Example: 'Max'

Data Types: double

Output Arguments

collapse all

Normalized intensity values, returned as a matrix.

Normalization parameters that you can use to normalize another group of signals, returned as a structure.

Version History

Introduced before R2006a