Comparative evaluation of set-level techniques in microarray classification

Jiri Klema, Matej Holec, Filip Zelezny, Jakub Tolar

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Analysis of gene expression data in terms of a priori-defined gene sets typically yields more compact and interpretable results than those produced by traditional methods that rely on individual genes. The set-level strategy can also be adopted in predictive classification tasks accomplished with machine learning algorithms. Here, sample features originally corresponding to genes are replaced by a much smaller number of features, each corresponding to a gene set and aggregating expressions of its members into a single real value. Classifiers learned from such transformed features promise better interpretability in that they derive class predictions from overall expressions of selected gene sets (e.g. corresponding to pathways) rather than expressions of specific genes. In a large collection of experiments we test how accurate such classifiers are compared to traditional classifiers based on genes. Furthermore, we translate some recently published gene set analysis techniques to the above proposed machine learning setting and assess their contributions to the classification accuracies.

Original languageEnglish (US)
Title of host publicationBioinformatics Research and Applications - 7th International Symposium, ISBRA 2011, Proceedings
Pages274-285
Number of pages12
DOIs
StatePublished - May 16 2011
Event7th International Symposium on Bioinformatics Research and Applications, ISBRA 2011 - Changsha, China
Duration: May 27 2011May 29 2011

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume6674 LNBI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other7th International Symposium on Bioinformatics Research and Applications, ISBRA 2011
CountryChina
CityChangsha
Period5/27/115/29/11

Fingerprint

Microarrays
Microarray
Level Set
Genes
Gene
Evaluation
Classifiers
Classifier
Learning systems
Machine Learning
Interpretability
Gene Expression Data
Gene expression
Learning algorithms
Learning Algorithm
Pathway
Prediction
Experiment

Keywords

  • classifer
  • gene set
  • learning
  • predictive accuracy

Cite this

Klema, J., Holec, M., Zelezny, F., & Tolar, J. (2011). Comparative evaluation of set-level techniques in microarray classification. In Bioinformatics Research and Applications - 7th International Symposium, ISBRA 2011, Proceedings (pp. 274-285). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6674 LNBI). https://doi.org/10.1007/978-3-642-21260-4_27

Comparative evaluation of set-level techniques in microarray classification. / Klema, Jiri; Holec, Matej; Zelezny, Filip; Tolar, Jakub.

Bioinformatics Research and Applications - 7th International Symposium, ISBRA 2011, Proceedings. 2011. p. 274-285 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6674 LNBI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Klema, J, Holec, M, Zelezny, F & Tolar, J 2011, Comparative evaluation of set-level techniques in microarray classification. in Bioinformatics Research and Applications - 7th International Symposium, ISBRA 2011, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 6674 LNBI, pp. 274-285, 7th International Symposium on Bioinformatics Research and Applications, ISBRA 2011, Changsha, China, 5/27/11. https://doi.org/10.1007/978-3-642-21260-4_27
Klema J, Holec M, Zelezny F, Tolar J. Comparative evaluation of set-level techniques in microarray classification. In Bioinformatics Research and Applications - 7th International Symposium, ISBRA 2011, Proceedings. 2011. p. 274-285. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-21260-4_27
Klema, Jiri ; Holec, Matej ; Zelezny, Filip ; Tolar, Jakub. / Comparative evaluation of set-level techniques in microarray classification. Bioinformatics Research and Applications - 7th International Symposium, ISBRA 2011, Proceedings. 2011. pp. 274-285 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{3112e271767b47b386075a49a014500f,
title = "Comparative evaluation of set-level techniques in microarray classification",
abstract = "Analysis of gene expression data in terms of a priori-defined gene sets typically yields more compact and interpretable results than those produced by traditional methods that rely on individual genes. The set-level strategy can also be adopted in predictive classification tasks accomplished with machine learning algorithms. Here, sample features originally corresponding to genes are replaced by a much smaller number of features, each corresponding to a gene set and aggregating expressions of its members into a single real value. Classifiers learned from such transformed features promise better interpretability in that they derive class predictions from overall expressions of selected gene sets (e.g. corresponding to pathways) rather than expressions of specific genes. In a large collection of experiments we test how accurate such classifiers are compared to traditional classifiers based on genes. Furthermore, we translate some recently published gene set analysis techniques to the above proposed machine learning setting and assess their contributions to the classification accuracies.",
keywords = "classifer, gene set, learning, predictive accuracy",
author = "Jiri Klema and Matej Holec and Filip Zelezny and Jakub Tolar",
year = "2011",
month = "5",
day = "16",
doi = "10.1007/978-3-642-21260-4_27",
language = "English (US)",
isbn = "9783642212598",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "274--285",
booktitle = "Bioinformatics Research and Applications - 7th International Symposium, ISBRA 2011, Proceedings",

}

TY - GEN

T1 - Comparative evaluation of set-level techniques in microarray classification

AU - Klema, Jiri

AU - Holec, Matej

AU - Zelezny, Filip

AU - Tolar, Jakub

PY - 2011/5/16

Y1 - 2011/5/16

N2 - Analysis of gene expression data in terms of a priori-defined gene sets typically yields more compact and interpretable results than those produced by traditional methods that rely on individual genes. The set-level strategy can also be adopted in predictive classification tasks accomplished with machine learning algorithms. Here, sample features originally corresponding to genes are replaced by a much smaller number of features, each corresponding to a gene set and aggregating expressions of its members into a single real value. Classifiers learned from such transformed features promise better interpretability in that they derive class predictions from overall expressions of selected gene sets (e.g. corresponding to pathways) rather than expressions of specific genes. In a large collection of experiments we test how accurate such classifiers are compared to traditional classifiers based on genes. Furthermore, we translate some recently published gene set analysis techniques to the above proposed machine learning setting and assess their contributions to the classification accuracies.

AB - Analysis of gene expression data in terms of a priori-defined gene sets typically yields more compact and interpretable results than those produced by traditional methods that rely on individual genes. The set-level strategy can also be adopted in predictive classification tasks accomplished with machine learning algorithms. Here, sample features originally corresponding to genes are replaced by a much smaller number of features, each corresponding to a gene set and aggregating expressions of its members into a single real value. Classifiers learned from such transformed features promise better interpretability in that they derive class predictions from overall expressions of selected gene sets (e.g. corresponding to pathways) rather than expressions of specific genes. In a large collection of experiments we test how accurate such classifiers are compared to traditional classifiers based on genes. Furthermore, we translate some recently published gene set analysis techniques to the above proposed machine learning setting and assess their contributions to the classification accuracies.

KW - classifer

KW - gene set

KW - learning

KW - predictive accuracy

UR - http://www.scopus.com/inward/record.url?scp=79955799256&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79955799256&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-21260-4_27

DO - 10.1007/978-3-642-21260-4_27

M3 - Conference contribution

SN - 9783642212598

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 274

EP - 285

BT - Bioinformatics Research and Applications - 7th International Symposium, ISBRA 2011, Proceedings

ER -