AbsoluteQuantification {aLFQ}

Absolute label-free quantification of mass spectrometry proteomics experiments

Description

Absolute label-free quantification of mass spectrometry proteomics experiments.

Usage

## Default S3 method:
AbsoluteQuantification(data, total_protein_concentration = 1, ...)
## S3 method for class 'AbsoluteQuantification'
cval(object, cval_method = "mc", mcx = 1000, ...)
## S3 method for class 'AbsoluteQuantification'
print(x, ...)
## S3 method for class 'AbsoluteQuantification'
plot(x, ...)
## S3 method for class 'AbsoluteQuantification'
hist(x, ...)
## S3 method for class 'AbsoluteQuantification'
pivot(x, ...)
## S3 method for class 'AbsoluteQuantification'
export(x, file, ...)

Arguments

data

a mandatory data frame containing the columns "run_id", "protein_id", "response", and "concentration" as generated by ProteinInference. The id column can be defined in any format, while the "response" and "concentration" columns need to be numeric and in non-log form. The data may contain calibration data (with numeric "concentration" and test data (with "concentration" = "?"))

total_protein_concentration

the total protein concentration in the sample in any unit. This will be used for the normalized protein and concentration columns.

object

an AbsoluteQuantification object.

cval_method

a method for doing crossvalidation: "boot" (bootstrapping), "mc" (monte carlo cross-validation), "loo" (leaving-one-out).

mcx

a positive integer value of the number of folds for cross-validation.

file

the location of the output csv file.

x

an AbsoluteQuantification object.

...

future extensions.

Details

If absolute quantity estimation based on anchor peptides or proteins is demanded, the calibration peptide or protein abundance must be provided. Both estimated calibration protein intensities and separately determined calibration protein concentrations are log transformed and a first order linear least-squares regression of this log-log data is calculated. The abundance of the target proteins is predicted based on this regression. The error of the regression arises from biological and technical variation as well from the protein and peptide intensity estimators. To perform model selection and to estimate the error of the predicted protein concentrations, bootstrapping and Monte Carlo cross-validation as suggested (Malmstrom et al., 2009; Ludwig et al., 2012) were implemented. For both methods, the objective function is the minimization of the mean fold-error.

If, on the other hand, the total protein concentration per cell is supplied in proteome-wide experiments, the absolute protein concentrations are estimated by normalization of the MS intensities or spectral counts to this number (Lu et al., 2006).

Value

An object of class AbsoluteQuantification.

Author(s)

George Rosenberger [e-mail]

References

Malmstrom, J. et al. Proteome-wide cellular protein concentrations of the human pathogen Leptospira interrogans. Nature 460, 762-765 (2009).

Ludwig, C., Claassen, M., Schmidt, A. & Aebersold, R. Estimation of Absolute Protein Quantities of Unlabeled Samples by Selected Reaction Monitoring Mass Spectrometry. Molecular & Cellular Proteomics 11, M111.013987-M111.013987 (2012).

Lu, P., Vogel, C., Wang, R., Yao, X. & Marcotte, E. M. Absolute protein expression profiling estimates the relative contributions of transcriptional and translational regulation. Nat Biotech 25, 117-124 (2006).

See Also

import, ProteinInference, ALF, APEX, apexFeatures, proteotypic

Examples

data(UPS2MS)

UPS2_SRM<-head(UPS2_SRM,100) # Remove this line for real applications
data_PI <- ProteinInference(UPS2_SRM)
data_AQ <- predict(cval(AbsoluteQuantification(data_PI),mcx=2))
print(data_AQ)
plot(data_AQ)
hist(data_AQ)
pivot(data_AQ)