Toward digital staining using imaging mass spectrometry and random forests

Michael Hanselmann; Ullrich Köthe; Marc Kirchner; Bernhard Y. Renard; Erika R. Amstalden; Kristine Glunde; Ron M.A. Heeren; Fred A. Hamprecht

doi:10.1021/pr900253y

Toward digital staining using imaging mass spectrometry and random forests

Michael Hanselmann, Ullrich Köthe, Marc Kirchner, Bernhard Y. Renard, Erika R. Amstalden, Kristine Glunde, Ron M.A. Heeren, Fred A. Hamprecht

School of Medicine

Research output: Contribution to journal › Article › peer-review

64 Scopus citations

Abstract

We show on imaging mass spectrometry (IMS) data that the Random Forest classifier can be used for automated tissue classification and that it results in predictions with high sensitivities and positive predictive values, even when intersample variability is present in the data. We further demonstrate how Markov Random Fields and vector-valued median filtering can be applied to reduce noise effects to further improve the classification results in a posthoc smoothing step. Our study gives clear evidence that digital staining by means of IMS constitutes a promising complement to chemical staining techniques.

Original language	English (US)
Pages (from-to)	3558-3567
Number of pages	10
Journal	Journal of proteome research
Volume	8
Issue number	7
DOIs	https://doi.org/10.1021/pr900253y
State	Published - Jul 6 2009

Keywords

Bioinformatics
Hyperspectral images
Imaging mass spectrometry
Markov Random Fields
Random Forest classification
Smoothing
Spectral images

ASJC Scopus subject areas

General Chemistry
Biochemistry

Access to Document

10.1021/pr900253y

Cite this

@article{8f49195d3fbe4e8c9bdd4809e6cebc4a,

title = "Toward digital staining using imaging mass spectrometry and random forests",

abstract = "We show on imaging mass spectrometry (IMS) data that the Random Forest classifier can be used for automated tissue classification and that it results in predictions with high sensitivities and positive predictive values, even when intersample variability is present in the data. We further demonstrate how Markov Random Fields and vector-valued median filtering can be applied to reduce noise effects to further improve the classification results in a posthoc smoothing step. Our study gives clear evidence that digital staining by means of IMS constitutes a promising complement to chemical staining techniques.",

keywords = "Bioinformatics, Hyperspectral images, Imaging mass spectrometry, Markov Random Fields, Random Forest classification, Smoothing, Spectral images",

author = "Michael Hanselmann and Ullrich K{\"o}the and Marc Kirchner and Renard, {Bernhard Y.} and Amstalden, {Erika R.} and Kristine Glunde and Heeren, {Ron M.A.} and Hamprecht, {Fred A.}",

year = "2009",

month = jul,

day = "6",

doi = "10.1021/pr900253y",

language = "English (US)",

volume = "8",

pages = "3558--3567",

journal = "Journal of proteome research",

issn = "1535-3893",

publisher = "American Chemical Society",

number = "7",

}

TY - JOUR

T1 - Toward digital staining using imaging mass spectrometry and random forests

AU - Hanselmann, Michael

AU - Köthe, Ullrich

AU - Kirchner, Marc

AU - Renard, Bernhard Y.

AU - Amstalden, Erika R.

AU - Glunde, Kristine

AU - Heeren, Ron M.A.

AU - Hamprecht, Fred A.

PY - 2009/7/6

Y1 - 2009/7/6

N2 - We show on imaging mass spectrometry (IMS) data that the Random Forest classifier can be used for automated tissue classification and that it results in predictions with high sensitivities and positive predictive values, even when intersample variability is present in the data. We further demonstrate how Markov Random Fields and vector-valued median filtering can be applied to reduce noise effects to further improve the classification results in a posthoc smoothing step. Our study gives clear evidence that digital staining by means of IMS constitutes a promising complement to chemical staining techniques.

AB - We show on imaging mass spectrometry (IMS) data that the Random Forest classifier can be used for automated tissue classification and that it results in predictions with high sensitivities and positive predictive values, even when intersample variability is present in the data. We further demonstrate how Markov Random Fields and vector-valued median filtering can be applied to reduce noise effects to further improve the classification results in a posthoc smoothing step. Our study gives clear evidence that digital staining by means of IMS constitutes a promising complement to chemical staining techniques.

KW - Bioinformatics

KW - Hyperspectral images

KW - Imaging mass spectrometry

KW - Markov Random Fields

KW - Random Forest classification

KW - Smoothing

KW - Spectral images

UR - http://www.scopus.com/inward/record.url?scp=67650370012&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67650370012&partnerID=8YFLogxK

U2 - 10.1021/pr900253y

DO - 10.1021/pr900253y

M3 - Article

C2 - 19469555

AN - SCOPUS:67650370012

SN - 1535-3893

VL - 8

SP - 3558

EP - 3567

JO - Journal of proteome research

JF - Journal of proteome research

IS - 7

ER -

Toward digital staining using imaging mass spectrometry and random forests

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this