Layerwise feature selection in Stacked Sparse Auto-Encoder for tumor type prediction

Vikas Singh, Nikhil Baranwal, Rahul K. Sevakula, Nishchal K. Verma, Yan Cui

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

6 Citations (Scopus)

Abstract

Transcriptome data has been proved to be very valuable for clinical applications, such as diagnosis and prognosis of various cancers. In this paper, we present layer-wise feature selection in conjunction with stacked sparse auto-encoders (SSAE), a deep learning strategy for tumor classification with gene expression data. While SSAE learns high-level features from data, performing feature selection in every layer is a heuristic to obtain relevant features at every stage and also to assist in reducing the computation during fine-tuning procedure. The data in the new feature representation is finally used by classifier(s) to perform Tumor detection. The algorithm was tested on 36 datasets from the GEMLeR repository and w.r.t. AUC (Area under ROC curve) performance, it was found to outperform the GEMLeR benchmark results on 35 datasets (tied on the other dataset).

Original languageEnglish (US)
Title of host publicationProceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016
EditorsKevin Burrage, Qian Zhu, Yunlong Liu, Tianhai Tian, Yadong Wang, Xiaohua Tony Hu, Qinghua Jiang, Jiangning Song, Shinichi Morishita, Kevin Burrage, Guohua Wang
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1542-1548
Number of pages7
ISBN (Electronic)9781509016105
DOIs
StatePublished - Jan 17 2017
Event2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016 - Shenzhen, China
Duration: Dec 15 2016Dec 18 2016

Publication series

NameProceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016

Other

Other2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016
CountryChina
CityShenzhen
Period12/15/1612/18/16

Fingerprint

Feature extraction
Tumors
Gene expression
Benchmarking
Neoplasms
Classifiers
Tuning
Transcriptome
ROC Curve
Area Under Curve
Learning
Gene Expression
Datasets
Deep learning
Heuristics

All Science Journal Classification (ASJC) codes

  • Genetics
  • Medicine (miscellaneous)
  • Genetics(clinical)
  • Biochemistry, medical
  • Biochemistry
  • Molecular Medicine
  • Health Informatics

Cite this

Singh, V., Baranwal, N., Sevakula, R. K., Verma, N. K., & Cui, Y. (2017). Layerwise feature selection in Stacked Sparse Auto-Encoder for tumor type prediction. In K. Burrage, Q. Zhu, Y. Liu, T. Tian, Y. Wang, X. T. Hu, Q. Jiang, J. Song, S. Morishita, K. Burrage, ... G. Wang (Eds.), Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016 (pp. 1542-1548). [7822750] (Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/BIBM.2016.7822750

Layerwise feature selection in Stacked Sparse Auto-Encoder for tumor type prediction. / Singh, Vikas; Baranwal, Nikhil; Sevakula, Rahul K.; Verma, Nishchal K.; Cui, Yan.

Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016. ed. / Kevin Burrage; Qian Zhu; Yunlong Liu; Tianhai Tian; Yadong Wang; Xiaohua Tony Hu; Qinghua Jiang; Jiangning Song; Shinichi Morishita; Kevin Burrage; Guohua Wang. Institute of Electrical and Electronics Engineers Inc., 2017. p. 1542-1548 7822750 (Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016).

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

Singh, V, Baranwal, N, Sevakula, RK, Verma, NK & Cui, Y 2017, Layerwise feature selection in Stacked Sparse Auto-Encoder for tumor type prediction. in K Burrage, Q Zhu, Y Liu, T Tian, Y Wang, XT Hu, Q Jiang, J Song, S Morishita, K Burrage & G Wang (eds), Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016., 7822750, Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016, Institute of Electrical and Electronics Engineers Inc., pp. 1542-1548, 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016, Shenzhen, China, 12/15/16. https://doi.org/10.1109/BIBM.2016.7822750
Singh V, Baranwal N, Sevakula RK, Verma NK, Cui Y. Layerwise feature selection in Stacked Sparse Auto-Encoder for tumor type prediction. In Burrage K, Zhu Q, Liu Y, Tian T, Wang Y, Hu XT, Jiang Q, Song J, Morishita S, Burrage K, Wang G, editors, Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016. Institute of Electrical and Electronics Engineers Inc. 2017. p. 1542-1548. 7822750. (Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016). https://doi.org/10.1109/BIBM.2016.7822750
Singh, Vikas ; Baranwal, Nikhil ; Sevakula, Rahul K. ; Verma, Nishchal K. ; Cui, Yan. / Layerwise feature selection in Stacked Sparse Auto-Encoder for tumor type prediction. Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016. editor / Kevin Burrage ; Qian Zhu ; Yunlong Liu ; Tianhai Tian ; Yadong Wang ; Xiaohua Tony Hu ; Qinghua Jiang ; Jiangning Song ; Shinichi Morishita ; Kevin Burrage ; Guohua Wang. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 1542-1548 (Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016).
@inproceedings{b9d5a3e6cd2a4dcf8681d8e35af75aeb,
title = "Layerwise feature selection in Stacked Sparse Auto-Encoder for tumor type prediction",
abstract = "Transcriptome data has been proved to be very valuable for clinical applications, such as diagnosis and prognosis of various cancers. In this paper, we present layer-wise feature selection in conjunction with stacked sparse auto-encoders (SSAE), a deep learning strategy for tumor classification with gene expression data. While SSAE learns high-level features from data, performing feature selection in every layer is a heuristic to obtain relevant features at every stage and also to assist in reducing the computation during fine-tuning procedure. The data in the new feature representation is finally used by classifier(s) to perform Tumor detection. The algorithm was tested on 36 datasets from the GEMLeR repository and w.r.t. AUC (Area under ROC curve) performance, it was found to outperform the GEMLeR benchmark results on 35 datasets (tied on the other dataset).",
author = "Vikas Singh and Nikhil Baranwal and Sevakula, {Rahul K.} and Verma, {Nishchal K.} and Yan Cui",
year = "2017",
month = "1",
day = "17",
doi = "10.1109/BIBM.2016.7822750",
language = "English (US)",
series = "Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "1542--1548",
editor = "Kevin Burrage and Qian Zhu and Yunlong Liu and Tianhai Tian and Yadong Wang and Hu, {Xiaohua Tony} and Qinghua Jiang and Jiangning Song and Shinichi Morishita and Kevin Burrage and Guohua Wang",
booktitle = "Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016",
address = "United States",

}

TY - GEN

T1 - Layerwise feature selection in Stacked Sparse Auto-Encoder for tumor type prediction

AU - Singh, Vikas

AU - Baranwal, Nikhil

AU - Sevakula, Rahul K.

AU - Verma, Nishchal K.

AU - Cui, Yan

PY - 2017/1/17

Y1 - 2017/1/17

N2 - Transcriptome data has been proved to be very valuable for clinical applications, such as diagnosis and prognosis of various cancers. In this paper, we present layer-wise feature selection in conjunction with stacked sparse auto-encoders (SSAE), a deep learning strategy for tumor classification with gene expression data. While SSAE learns high-level features from data, performing feature selection in every layer is a heuristic to obtain relevant features at every stage and also to assist in reducing the computation during fine-tuning procedure. The data in the new feature representation is finally used by classifier(s) to perform Tumor detection. The algorithm was tested on 36 datasets from the GEMLeR repository and w.r.t. AUC (Area under ROC curve) performance, it was found to outperform the GEMLeR benchmark results on 35 datasets (tied on the other dataset).

AB - Transcriptome data has been proved to be very valuable for clinical applications, such as diagnosis and prognosis of various cancers. In this paper, we present layer-wise feature selection in conjunction with stacked sparse auto-encoders (SSAE), a deep learning strategy for tumor classification with gene expression data. While SSAE learns high-level features from data, performing feature selection in every layer is a heuristic to obtain relevant features at every stage and also to assist in reducing the computation during fine-tuning procedure. The data in the new feature representation is finally used by classifier(s) to perform Tumor detection. The algorithm was tested on 36 datasets from the GEMLeR repository and w.r.t. AUC (Area under ROC curve) performance, it was found to outperform the GEMLeR benchmark results on 35 datasets (tied on the other dataset).

UR - http://www.scopus.com/inward/record.url?scp=85013336347&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85013336347&partnerID=8YFLogxK

U2 - 10.1109/BIBM.2016.7822750

DO - 10.1109/BIBM.2016.7822750

M3 - Conference contribution

T3 - Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016

SP - 1542

EP - 1548

BT - Proceedings - 2016 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2016

A2 - Burrage, Kevin

A2 - Zhu, Qian

A2 - Liu, Yunlong

A2 - Tian, Tianhai

A2 - Wang, Yadong

A2 - Hu, Xiaohua Tony

A2 - Jiang, Qinghua

A2 - Song, Jiangning

A2 - Morishita, Shinichi

A2 - Burrage, Kevin

A2 - Wang, Guohua

PB - Institute of Electrical and Electronics Engineers Inc.

ER -