CBioProfiler

CBioProfiler includes two pipelines for analyzing cancer biomarkers and subtypes using multiple survival-related machine learning algorithms and unsupervised machine learning methods, integrated with curated cancer prognosis data from 268 studies.

Key Features:

• Cancer Biomarker Pipeline: Identifies and annotates cancer survival-related biomarkers using multiple survival-related machine learning algorithms across five modules.
• Cancer Subtype Pipeline: Performs data preprocessing, subtype identification using multiple unsupervised machine learning methods, and subtype evaluation and validation across three modules.
• CuratedCancerPrognosisData: Integrates reviewed and curated gene expression and clinical data from 268 studies covering 43 common blood and solid tumors and 47,686 clinical samples.

Scientific Applications:

• Cancer Biomarker Identification: Identifies survival-related biomarkers for risk stratification, treatment, and prognosis prediction in cancer patients.
• Cancer Subtype Characterization: Identifies and validates cancer subtypes using unsupervised machine learning methods for patient classification.

Methodology:

The cancer biomarker pipeline uses multiple survival-related machine learning algorithms. The cancer subtype pipeline uses multiple unsupervised machine learning methods for subtype identification, followed by evaluation and validation.