Skip to Main content Skip to Navigation
Conference papers

Unsupervised feature selection for tumor profiles using autoencoders and kernel methods

Abstract : Molecular data from tumor profiles is high dimensional. Tumor profiles can be characterized by tens of thousands of gene expression features. Due to the size of the gene expression feature set machine learning methods are exposed to noisy variables and complexity. Tumor types present heterogeneity and can be subdivided in tumor subtypes. In many cases tumor data does not include tumor subtype labeling thus unsupervised learning methods are necessary for tumor subtype discovery. This work aims to learn meaningful and low dimensional representations of tumor samples and find tumor subtype clusters while keeping biological signatures without using tumor labels. By using Autoencoders a low dimensional and denoised latent space is learned as a target representation to guide a Multiple Kernel Learning model that selects a subset of genes. By using the selected genes a clustering method is used to group samples. In order to evaluate the performance of the proposed unsupervised feature selection method the obtained features and clusters are analyzed by clinical significance. The proposed method has been applied on three tumor datasets which are Brain, Renal and Lung, each one composed by two tumor subtypes. The results obtained by the proposed method reveal lower redundancy in the selected features and each obtained cluster is significantly enriched by just one tumor subtype when compared with benchmark unsupervised feature selection methods. The proposed method named Latent Kernel Feature Selection (LKFS) is an unsupervised approach for gene selection in tumor gene expression profiles.
Document type :
Conference papers
Complete list of metadata

https://hal-utt.archives-ouvertes.fr/hal-03320947
Contributor : Jean-Baptiste Vu Van Connect in order to contact the contributor
Submitted on : Monday, August 16, 2021 - 4:59:53 PM
Last modification on : Friday, August 27, 2021 - 3:14:07 PM

Links full text

Identifiers

Collections

UTT | CNRS

Citation

Martin Palazzo, Pierre Beauseroy, Patricio Yankilevich. Unsupervised feature selection for tumor profiles using autoencoders and kernel methods. 2020 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB), Oct 2020, Via del Mar, Chile. pp.1-8, ⟨10.1109/CIBCB48159.2020.9277699⟩. ⟨hal-03320947⟩

Share

Metrics

Record views

20