Next Article in Journal
Aberrant Sialylation in Cancer: Biomarker and Potential Target for Therapeutic Intervention?
Next Article in Special Issue
Prediction of Incident Cancers in the Lifelines Population-Based Cohort
Previous Article in Journal
Inhibition of the Myocardin-Related Transcription Factor Pathway Increases Efficacy of Trametinib in NRAS-Mutant Melanoma Cell Lines
Previous Article in Special Issue
Assessment of Immunological Features in Muscle-Invasive Bladder Cancer Prognosis Using Ensemble Learning
Article

Performance Comparison of Deep Learning Autoencoders for Cancer Subtype Detection Using Multi-Omics Data

1
Institute of Biological Sciences, Federal University of Para, Belem, PA 66075-110, Brazil
2
Laboratory of Virology and Environmental Genomics, Instituto de Innovacion en Biotecnologia e Industria (IIBI), Santo Domingo 10104, Dominican Republic
3
Instituto Tecnológico de Santo Domingo (INTEC), Santo Domingo 10602, Dominican Republic
4
Department of Computer Science, Virginia Commonwealth University, Richmond, VA 23284, USA
5
Programa de Pós-Graduação em Enfermagem, Federal University of Para, Belem, PA 66075-110, Brazil
6
Institute of Biological Science, Federal University of Minas Gerais, Belo Horizonte, MG 31270-901, Brazil
*
Author to whom correspondence should be addressed.
These authors contributed equally.
Academic Editor: Ognjen Arandjelović
Cancers 2021, 13(9), 2013; https://doi.org/10.3390/cancers13092013
Received: 15 February 2021 / Revised: 29 March 2021 / Accepted: 6 April 2021 / Published: 22 April 2021
(This article belongs to the Special Issue Machine Learning Techniques in Cancer)
Here, we compared the performance of four different autoencoders: (a) vanilla, (b) sparse, (c) denoising, and (d) variational for subtype detection on four cancer types: Glioblastoma multiforme, Colon Adenocarcinoma, Kidney renal clear cell carcinoma, and Breast invasive carcinoma. Multiview dataset comprising gene expression, DNA methylation, and miRNA expression from TCGA is fed into an autoencoder to get a compressed nonlinear representation. Then the clustering technique was applied on that compressed representation to reveal the subtype of cancer. Though different autoencoders’ performance varies on different datasets, they performed much better than standard data fusion techniques such as PCA, kernel PCA, and sparse PCA.
A heterogeneous disease such as cancer is activated through multiple pathways and different perturbations. Depending upon the activated pathway(s), the survival of the patients varies significantly and shows different efficacy to various drugs. Therefore, cancer subtype detection using genomics level data is a significant research problem. Subtype detection is often a complex problem, and in most cases, needs multi-omics data fusion to achieve accurate subtyping. Different data fusion and subtyping approaches have been proposed over the years, such as kernel-based fusion, matrix factorization, and deep learning autoencoders. In this paper, we compared the performance of different deep learning autoencoders for cancer subtype detection. We performed cancer subtype detection on four different cancer types from The Cancer Genome Atlas (TCGA) datasets using four autoencoder implementations. We also predicted the optimal number of subtypes in a cancer type using the silhouette score and found that the detected subtypes exhibit significant differences in survival profiles. Furthermore, we compared the effect of feature selection and similarity measures for subtype detection. For further evaluation, we used the Glioblastoma multiforme (GBM) dataset and identified the differentially expressed genes in each of the subtypes. The results obtained are consistent with other genomic studies and can be corroborated with the involved pathways and biological functions. Thus, it shows that the results from the autoencoders, obtained through the interaction of different datatypes of cancer, can be used for the prediction and characterization of patient subgroups and survival profiles. View Full-Text
Keywords: cancer subtype detection; multi-omics data; data integration; autoencoder; survival analysis cancer subtype detection; multi-omics data; data integration; autoencoder; survival analysis
Show Figures

Figure 1

  • Supplementary File 1:

    ZIP-Document (ZIP, 23152 KiB)

  • Externally hosted supplementary file 1
    Doi: 10.5281/zenodo.4540843
    Link: https://doi.org/10.5281/zenodo.4540843
    Description: Supplementary data 1: It contains the images of the results of the survival analyses obtained by integrating the data carried out by the different autoencoder algorithms. Supplementary data 1: This contains the differential expression analysis results of the different clusters identified in the GBM dataset.
MDPI and ACS Style

Franco, E.F.; Rana, P.; Cruz, A.; Calderón, V.V.; Azevedo, V.; Ramos, R.T.J.; Ghosh, P. Performance Comparison of Deep Learning Autoencoders for Cancer Subtype Detection Using Multi-Omics Data. Cancers 2021, 13, 2013. https://doi.org/10.3390/cancers13092013

AMA Style

Franco EF, Rana P, Cruz A, Calderón VV, Azevedo V, Ramos RTJ, Ghosh P. Performance Comparison of Deep Learning Autoencoders for Cancer Subtype Detection Using Multi-Omics Data. Cancers. 2021; 13(9):2013. https://doi.org/10.3390/cancers13092013

Chicago/Turabian Style

Franco, Edian F., Pratip Rana, Aline Cruz, Víctor V. Calderón, Vasco Azevedo, Rommel T.J. Ramos, and Preetam Ghosh. 2021. "Performance Comparison of Deep Learning Autoencoders for Cancer Subtype Detection Using Multi-Omics Data" Cancers 13, no. 9: 2013. https://doi.org/10.3390/cancers13092013

Find Other Styles
Note that from the first issue of 2016, MDPI journals use article numbers instead of page numbers. See further details here.

Article Access Map by Country/Region

1
Back to TopTop