Cox-PASNet

Cox-PASNet is a biologically interpretable pathway-based sparse deep neural network for survival analysis of cancer patients using high-dimensional genomic and clinical data. The key features and advantages of Cox-PASNet are:

1. Interpretability: The nodes in the neural network correspond to biological genes and pathways, making the model biologically interpretable.

2. Nonlinear and hierarchical effects: Cox-PASNet captures the nonlinear and hierarchical effects of biological pathways associated with cancer patient survival.

3. Integration of data: The model integrates high-dimensional gene expression data and clinical data on a simple neural network architecture.

4. Handling HDLSS data: A heuristic optimization solution is proposed to train Cox-PASNet with high-dimension, low-sample size (HDLSS) data, which is common in genomic studies.

6. Identification of prognostic factors: The neural network architecture of Cox-PASNet can identify significant prognostic factors of genes and biological pathways.

Topic

Machine learning;Molecular interactions, pathways and networks;Oncology;Biomarkers;Preclinical and clinical studies

Detail

  • Operation: Validation;Pathway or network analysis;Expression analysis

  • Software interface: Command-line user interface

  • Language: Python

  • License: Not stated

  • Cost: Free of charge

  • Version name: -

  • Credit: National Institutes of Health/National Cancer Institute, University of Nevada, Las Vegas, Ministry of Science, ICT, Korea, Howard R. Hughes College of Engineering.

  • Input: -

  • Output: -

  • Contact: Mingon Kang mingon.kang@unlv.edu

  • Collection: -

  • Maturity: -

Publications

  • Interpretable deep neural network for cancer survival analysis by integrating genomic and clinical data.
  • Hao J, et al. Interpretable deep neural network for cancer survival analysis by integrating genomic and clinical data. Interpretable deep neural network for cancer survival analysis by integrating genomic and clinical data. 2019; 12:189. doi: 10.1186/s12920-019-0624-2
  • https://doi.org/10.1186/S12920-019-0624-2
  • PMID: 31865908
  • PMC: PMC6927105

Download and documentation

< Back to DB search