SVM model for quality assessment of medium resolution mass spectra from 18O-water labeling experiments

Alexey V. Nefedov, Miroslaw J. Gilski, Rovshan Sadygov

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

We describe a method for assessing the quality of mass spectra and improving reliability of relative ratio estimations from 18O-water labeling experiments acquired from low resolution mass spectrometers. The mass profiles of heavy and light peptide pairs are often affected by artifacts, including coeluting contaminant species, noise signal, instrumental fluctuations in measuring ion position and abundance levels. Such artifacts distort the profiles, leading to erroneous ratio estimations, thus reducing the reliability of ratio estimations in high throughput quantification experiments. We used support vector machines (SVMs) to filter out mass spectra that deviated significantly from expected theoretical isotope distributions. We built an SVM classifier with a decision function that assigns a score to every mass profile based on such spectral features as mass accuracy, signal-to-noise ratio, and differences between experimental and theoretical isotopic distributions. The classifier was trained using a data set obtained from samples of mouse renal cortex. We then tested it on protein samples (bovine serum albumin) mixed in five different ratios of labeled and unlabeled species. We demonstrated that filtering the data using our SVM classifier results in as much as a 9-fold reduction in the coefficient of variance of peptide ratios, thus significantly improving the reliability of ratio estimations.

Original languageEnglish (US)
Pages (from-to)2095-2103
Number of pages9
JournalJournal of Proteome Research
Volume10
Issue number4
DOIs
StatePublished - Apr 1 2011

Fingerprint

Mass Media
Labeling
Support vector machines
Artifacts
Classifiers
Water
Peptides
Experiments
Signal-To-Noise Ratio
Bovine Serum Albumin
Isotopes
Noise
Mass spectrometers
Ions
Kidney
Light
Signal to noise ratio
Throughput
Impurities
Support Vector Machine

Keywords

  • isotope distribution
  • mass accuracy
  • signal-to-noise ratio
  • stable-isotope labeling
  • Support vector machines

ASJC Scopus subject areas

  • Biochemistry
  • Chemistry(all)

Cite this

SVM model for quality assessment of medium resolution mass spectra from 18O-water labeling experiments. / Nefedov, Alexey V.; Gilski, Miroslaw J.; Sadygov, Rovshan.

In: Journal of Proteome Research, Vol. 10, No. 4, 01.04.2011, p. 2095-2103.

Research output: Contribution to journalArticle

@article{47bc3f72781645c9bc138b62973358b1,
title = "SVM model for quality assessment of medium resolution mass spectra from 18O-water labeling experiments",
abstract = "We describe a method for assessing the quality of mass spectra and improving reliability of relative ratio estimations from 18O-water labeling experiments acquired from low resolution mass spectrometers. The mass profiles of heavy and light peptide pairs are often affected by artifacts, including coeluting contaminant species, noise signal, instrumental fluctuations in measuring ion position and abundance levels. Such artifacts distort the profiles, leading to erroneous ratio estimations, thus reducing the reliability of ratio estimations in high throughput quantification experiments. We used support vector machines (SVMs) to filter out mass spectra that deviated significantly from expected theoretical isotope distributions. We built an SVM classifier with a decision function that assigns a score to every mass profile based on such spectral features as mass accuracy, signal-to-noise ratio, and differences between experimental and theoretical isotopic distributions. The classifier was trained using a data set obtained from samples of mouse renal cortex. We then tested it on protein samples (bovine serum albumin) mixed in five different ratios of labeled and unlabeled species. We demonstrated that filtering the data using our SVM classifier results in as much as a 9-fold reduction in the coefficient of variance of peptide ratios, thus significantly improving the reliability of ratio estimations.",
keywords = "isotope distribution, mass accuracy, signal-to-noise ratio, stable-isotope labeling, Support vector machines",
author = "Nefedov, {Alexey V.} and Gilski, {Miroslaw J.} and Rovshan Sadygov",
year = "2011",
month = "4",
day = "1",
doi = "10.1021/pr1012174",
language = "English (US)",
volume = "10",
pages = "2095--2103",
journal = "Journal of Proteome Research",
issn = "1535-3893",
publisher = "American Chemical Society",
number = "4",

}

TY - JOUR

T1 - SVM model for quality assessment of medium resolution mass spectra from 18O-water labeling experiments

AU - Nefedov, Alexey V.

AU - Gilski, Miroslaw J.

AU - Sadygov, Rovshan

PY - 2011/4/1

Y1 - 2011/4/1

N2 - We describe a method for assessing the quality of mass spectra and improving reliability of relative ratio estimations from 18O-water labeling experiments acquired from low resolution mass spectrometers. The mass profiles of heavy and light peptide pairs are often affected by artifacts, including coeluting contaminant species, noise signal, instrumental fluctuations in measuring ion position and abundance levels. Such artifacts distort the profiles, leading to erroneous ratio estimations, thus reducing the reliability of ratio estimations in high throughput quantification experiments. We used support vector machines (SVMs) to filter out mass spectra that deviated significantly from expected theoretical isotope distributions. We built an SVM classifier with a decision function that assigns a score to every mass profile based on such spectral features as mass accuracy, signal-to-noise ratio, and differences between experimental and theoretical isotopic distributions. The classifier was trained using a data set obtained from samples of mouse renal cortex. We then tested it on protein samples (bovine serum albumin) mixed in five different ratios of labeled and unlabeled species. We demonstrated that filtering the data using our SVM classifier results in as much as a 9-fold reduction in the coefficient of variance of peptide ratios, thus significantly improving the reliability of ratio estimations.

AB - We describe a method for assessing the quality of mass spectra and improving reliability of relative ratio estimations from 18O-water labeling experiments acquired from low resolution mass spectrometers. The mass profiles of heavy and light peptide pairs are often affected by artifacts, including coeluting contaminant species, noise signal, instrumental fluctuations in measuring ion position and abundance levels. Such artifacts distort the profiles, leading to erroneous ratio estimations, thus reducing the reliability of ratio estimations in high throughput quantification experiments. We used support vector machines (SVMs) to filter out mass spectra that deviated significantly from expected theoretical isotope distributions. We built an SVM classifier with a decision function that assigns a score to every mass profile based on such spectral features as mass accuracy, signal-to-noise ratio, and differences between experimental and theoretical isotopic distributions. The classifier was trained using a data set obtained from samples of mouse renal cortex. We then tested it on protein samples (bovine serum albumin) mixed in five different ratios of labeled and unlabeled species. We demonstrated that filtering the data using our SVM classifier results in as much as a 9-fold reduction in the coefficient of variance of peptide ratios, thus significantly improving the reliability of ratio estimations.

KW - isotope distribution

KW - mass accuracy

KW - signal-to-noise ratio

KW - stable-isotope labeling

KW - Support vector machines

UR - http://www.scopus.com/inward/record.url?scp=79953696551&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79953696551&partnerID=8YFLogxK

U2 - 10.1021/pr1012174

DO - 10.1021/pr1012174

M3 - Article

VL - 10

SP - 2095

EP - 2103

JO - Journal of Proteome Research

JF - Journal of Proteome Research

SN - 1535-3893

IS - 4

ER -