cancer_subtyping

The software tool "cancer_subtyping" is an unsupervised deep learning method integrating multi-platform genomic data to identify subtypes of clear renal cell carcinoma (ccRCC). The technique uses stacked denoising autoencoders (SdA) to analyze five types of genomic data from The Cancer Genome Atlas (TCGA) and successfully identifies two distinct subtypes of ccRCC patients. The subtypes are associated with significant differences in survival probability, tumor grade, and pathological stage. The tool also identifies differentially expressed genes, proteins, and miRNAs between the two subtypes, with functional annotations matching the clinical features.

The pre-trained model learned from the Kidney Renal Clear Cell Carcinoma (KIRC) dataset can effectively be applied to other cancer types, such as Lung Adenocarcinoma (LUAD) and Low-Grade Glioma (LGG), to stratify patients into clinically relevant subtypes. The source code and models are available on GitHub to facilitate similar studies and assist in cancer subtyping, diagnosis, personalized treatments, and prognosis.

Topic

Oncology;Urology and nephrology;Machine learning;Functional, regulatory and non-coding RNA;Biomarkers

Detail

  • Operation: Expression correlation analysis;Gene-set enrichment analysis;miRNA expression analysis

  • Software interface: Command-line user interface

  • Language: Python

  • License: Not stated

  • Cost: Free of charge

  • Version name: -

  • Credit: The National Institute of Health, specifically the National Human Genome Research Institute (NHGRI) and National Cancer Institute (NCI).

  • Input: -

  • Output: -

  • Contact: Tongjun Gu tgu@ufl.edu ,Xiwu Zhao xiwuzhao@umich.edu

  • Collection: -

  • Maturity: -

Publications

  • Integrating multi-platform genomic datasets for kidney renal clear cell carcinoma subtyping using stacked denoising autoencoders.
  • Gu T and Zhao X. Integrating multi-platform genomic datasets for kidney renal clear cell carcinoma subtyping using stacked denoising autoencoders. Integrating multi-platform genomic datasets for kidney renal clear cell carcinoma subtyping using stacked denoising autoencoders. 2019; 9:16668. doi: 10.1038/s41598-019-53048-x
  • https://doi.org/10.1038/S41598-019-53048-X
  • PMID: 31723226
  • PMC: PMC6853929

Download and documentation

< Back to DB search