Estimating the statistical significance of classifiers by varying the number of genes

R. Maglietta; A. Piepoli; A. D'Addabbo; R. Cotugno; G. Pesole; S. Liuni; M. Savino; M. Carella; F. Perri; N. Ancona

doi:10.1109/GENSIPS.2006.353180

Estimating the statistical significance of classifiers by varying the number of genes

R. Maglietta, A. Piepoli, A. D'Addabbo, R. Cotugno, G. Pesole, S. Liuni, M. Savino, M. Carella, F. Perri, N. Ancona

IRCCS Casa Sollievo della Sofferenza

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We present a statistically well founded method to construct cancer predictors using gene expression profiles. This methodology is applied to a new microarray data set extracted from 25 patients affected by colon cancer. In particular, we answer to precise questions: how many gene expression levels are correlated with the pathology and how many are sufficient for an accurate classification? The proposed method provides answer to these questions avoiding the potential pitfalls hidden in the analysis of microarray data. We have evaluated the generalization error, estimated through the Leave-K-Out Cross Validation error, of two different classification schemes by varying the number of selected genes. We found that, Regularized Least Squares (RLS) and Support Vector Machines (SVM) classifiers, using the whole gene set, have error rates of e = 14% (p = 0.023) and e = 11% (p = 0.016) respectively. Concerning the number of genes, the performances of RLS and SVM classifiers do not change when the 74% of genes is used. The statistical significance was measured by using permutation test.

Original language	English
Title of host publication	2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006
Pages	109-110
Number of pages	2
DOIs	https://doi.org/10.1109/GENSIPS.2006.353180
Publication status	Published - 2006
Event	2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006 - College Station, TX, United States Duration: May 28 2006 → May 30 2006

Other

Other	2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006
Country	United States
City	College Station, TX
Period	5/28/06 → 5/30/06

ASJC Scopus subject areas

Biochemistry, Genetics and Molecular Biology (miscellaneous)
Computational Theory and Mathematics
Computer Vision and Pattern Recognition
Statistics and Probability

Access to Document

10.1109/GENSIPS.2006.353180

Fingerprint Dive into the research topics of 'Estimating the statistical significance of classifiers by varying the number of genes'. Together they form a unique fingerprint.

Cite this

Maglietta, R., Piepoli, A., D'Addabbo, A., Cotugno, R., Pesole, G., Liuni, S., Savino, M., Carella, M., Perri, F., & Ancona, N. (2006). Estimating the statistical significance of classifiers by varying the number of genes. In 2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006 (pp. 109-110). [4161801] https://doi.org/10.1109/GENSIPS.2006.353180

Estimating the statistical significance of classifiers by varying the number of genes. / Maglietta, R.; Piepoli, A.; D'Addabbo, A.; Cotugno, R.; Pesole, G.; Liuni, S.; Savino, M.; Carella, M.; Perri, F.; Ancona, N.

2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006. 2006. p. 109-110 4161801.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Maglietta, R, Piepoli, A, D'Addabbo, A, Cotugno, R, Pesole, G, Liuni, S, Savino, M, Carella, M , Perri, F & Ancona, N 2006, Estimating the statistical significance of classifiers by varying the number of genes. in 2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006., 4161801, pp. 109-110, 2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006, College Station, TX, United States, 5/28/06. https://doi.org/10.1109/GENSIPS.2006.353180

@inproceedings{589f2fd22a854bc281cfafdcd9653d87,

title = "Estimating the statistical significance of classifiers by varying the number of genes",

abstract = "We present a statistically well founded method to construct cancer predictors using gene expression profiles. This methodology is applied to a new microarray data set extracted from 25 patients affected by colon cancer. In particular, we answer to precise questions: how many gene expression levels are correlated with the pathology and how many are sufficient for an accurate classification? The proposed method provides answer to these questions avoiding the potential pitfalls hidden in the analysis of microarray data. We have evaluated the generalization error, estimated through the Leave-K-Out Cross Validation error, of two different classification schemes by varying the number of selected genes. We found that, Regularized Least Squares (RLS) and Support Vector Machines (SVM) classifiers, using the whole gene set, have error rates of e = 14% (p = 0.023) and e = 11% (p = 0.016) respectively. Concerning the number of genes, the performances of RLS and SVM classifiers do not change when the 74% of genes is used. The statistical significance was measured by using permutation test.",

author = "R. Maglietta and A. Piepoli and A. D'Addabbo and R. Cotugno and G. Pesole and S. Liuni and M. Savino and M. Carella and F. Perri and N. Ancona",

year = "2006",

doi = "10.1109/GENSIPS.2006.353180",

language = "English",

isbn = "1424403855",

pages = "109--110",

booktitle = "2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006",

note = "2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006 ; Conference date: 28-05-2006 Through 30-05-2006",

}

TY - GEN

T1 - Estimating the statistical significance of classifiers by varying the number of genes

AU - Maglietta, R.

AU - Piepoli, A.

AU - D'Addabbo, A.

AU - Cotugno, R.

AU - Pesole, G.

AU - Liuni, S.

AU - Savino, M.

AU - Carella, M.

AU - Perri, F.

AU - Ancona, N.

PY - 2006

Y1 - 2006

N2 - We present a statistically well founded method to construct cancer predictors using gene expression profiles. This methodology is applied to a new microarray data set extracted from 25 patients affected by colon cancer. In particular, we answer to precise questions: how many gene expression levels are correlated with the pathology and how many are sufficient for an accurate classification? The proposed method provides answer to these questions avoiding the potential pitfalls hidden in the analysis of microarray data. We have evaluated the generalization error, estimated through the Leave-K-Out Cross Validation error, of two different classification schemes by varying the number of selected genes. We found that, Regularized Least Squares (RLS) and Support Vector Machines (SVM) classifiers, using the whole gene set, have error rates of e = 14% (p = 0.023) and e = 11% (p = 0.016) respectively. Concerning the number of genes, the performances of RLS and SVM classifiers do not change when the 74% of genes is used. The statistical significance was measured by using permutation test.

AB - We present a statistically well founded method to construct cancer predictors using gene expression profiles. This methodology is applied to a new microarray data set extracted from 25 patients affected by colon cancer. In particular, we answer to precise questions: how many gene expression levels are correlated with the pathology and how many are sufficient for an accurate classification? The proposed method provides answer to these questions avoiding the potential pitfalls hidden in the analysis of microarray data. We have evaluated the generalization error, estimated through the Leave-K-Out Cross Validation error, of two different classification schemes by varying the number of selected genes. We found that, Regularized Least Squares (RLS) and Support Vector Machines (SVM) classifiers, using the whole gene set, have error rates of e = 14% (p = 0.023) and e = 11% (p = 0.016) respectively. Concerning the number of genes, the performances of RLS and SVM classifiers do not change when the 74% of genes is used. The statistical significance was measured by using permutation test.

UR - http://www.scopus.com/inward/record.url?scp=48649094295&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=48649094295&partnerID=8YFLogxK

U2 - 10.1109/GENSIPS.2006.353180

DO - 10.1109/GENSIPS.2006.353180

M3 - Conference contribution

AN - SCOPUS:48649094295

SN - 1424403855

SN - 9781424403851

SP - 109

EP - 110

BT - 2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006

T2 - 2006 IEEE International Workshop on Genomic Signal Processing and Statistics, GENSIPS 2006

Y2 - 28 May 2006 through 30 May 2006

ER -