Related Experiment Video
Updated: Jul 20, 2025
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
Deep multi-view contrastive learning for cancer subtype identification
Wenlan Chen1, Hong Wang1, Cheng Liang1
1School of Information Science and Engineering, Shandong Normal University, Jinan, 250358, China.
Abstract:
Cancer heterogeneity has posed great challenges in exploring precise therapeutic strategies for cancer treatment. The identification of cancer subtypes aims to detect patients with distinct molecular profiles and thus could provide new clues on effective clinical therapies. While great efforts have been made, it remains challenging to develop powerful computational methods that can efficiently integrate multi-omics datasets for the task. In this paper, we propose a novel self-supervised learning model called Deep Multi-view Contrastive Learning (DMCL) for cancer subtype identification. Specifically, by incorporating the reconstruction loss, contrastive loss and clustering loss into a unified framework, our model simultaneously encodes the sample discriminative information into the extracted feature representations and well preserves the sample cluster structures in the embedded space. Moreover, DMCL is an end-to-end framework where the cancer subtypes could be directly obtained from the model outputs. We compare DMCL with eight alternatives ranging from classic cancer subtype identification methods to recently developed state-of-the-art systems on 10 widely used cancer multi-omics datasets as well as an integrated dataset, and the experimental results validate the superior performance of our method. We further conduct a case study on liver cancer and the analysis results indicate that different subtypes might have different responses to the selected chemotherapeutic drugs.
Insights
Deep Multi-view Contrastive Learning (DMCL) effectively identifies cancer subtypes from multi-omics data. This approach aids in developing precise cancer therapies by revealing distinct molecular profiles and potential drug responses.
Area of Science:
- Computational biology
- Bioinformatics
- Machine learning in oncology
Background:
- Cancer heterogeneity presents significant challenges for developing precise therapeutic strategies.
- Identifying distinct cancer subtypes based on molecular profiles is crucial for effective clinical treatments.
- Integrating multi-omics datasets for cancer subtype identification requires advanced computational methods.
Purpose of the Study:
- To propose a novel self-supervised learning model, Deep Multi-view Contrastive Learning (DMCL), for accurate cancer subtype identification.
- To develop an end-to-end framework that integrates reconstruction, contrastive, and clustering losses for feature representation and cluster preservation.
- To demonstrate the capability of DMCL in directly outputting cancer subtypes.
Main Methods:
- Developed Deep Multi-view Contrastive Learning (DMCL), a self-supervised learning model.
- Incorporated reconstruction loss, contrastive loss, and clustering loss into a unified framework.
- Evaluated DMCL on 10 cancer multi-omics datasets and one integrated dataset, comparing against eight alternative methods.
Main Results:
- DMCL demonstrated superior performance compared to existing methods in cancer subtype identification across multiple datasets.
- The model effectively encodes sample discriminative information and preserves cluster structures in embedded feature representations.
- A case study on liver cancer indicated that identified subtypes may exhibit differential responses to chemotherapeutic drugs.
Conclusions:
- DMCL offers a powerful and efficient computational method for integrating multi-omics data for cancer subtype identification.
- The findings suggest DMCL's potential to guide personalized cancer treatment strategies by revealing subtype-specific drug sensitivities.
- This approach advances the field of precision oncology through improved subtype discovery from complex molecular data.
More Related Videos
07:15Machine Learning Algorithms for Early Detection of Bone Metastases in an Experimental Rat Model
Published on: August 16, 2020
04:09Predicting Treatment Response to Image-Guided Therapies Using Machine Learning: An Example for Trans-Arterial Treatment of Hepatocellular Carcinoma
Published on: October 10, 2018
Related Concept Videos
02:43Mouse Models of Cancer Study
The development of transgenic, knockout, and knock-in mice has led to an exponential increase in their use as model organisms in research,...
01:21Cancer Survival Analysis