Related Experiment Video
Updated: Jun 8, 2026
08:08
Quantitative Mass Spectrometric Profiling of Cancer-cell Proteomes Derived From Liquid and Solid Tumors
Published on: February 27, 2015
Semi-supervised recursively partitioned mixture models for identifying cancer subtypes.
Devin C Koestler1, Carmen J Marsit, Brock C Christensen
1Department of Community Health, Section for Biostatistics, Center for Environmental Health and Technology, Brown University, Providence, RI 02912, USA. devin_koestler@brown.edu
Bioinformatics (Oxford, England)
|September 14, 2010
Summary
We developed a new semi-supervised method to identify cancer subtypes linked to patient survival using genetic and clinical data. This approach helps understand disease progression and personalize cancer treatments.
Area of Science:
- Bioinformatics
- Computational Biology
- Cancer Genomics
Background:
- Cancer progression and treatment response vary significantly among patients with identical diagnoses.
- Histologically similar cancers can exhibit distinct molecular profiles, necessitating advanced analytical methods.
- Identifying molecular subtypes associated with patient survival is crucial for understanding disease mechanisms and developing personalized therapies.
Purpose of the Study:
- To introduce a novel semi-supervised method for identifying cancer subtypes associated with patient survival.
- To leverage array-based genetic and patient-level clinical data for subtype discovery.
- To provide a foundation for more personalized cancer treatment strategies.
Main Methods:
- Proposed semi-supervised recursively partitioned mixture models (SS-RPMM) utilizing genetic and clinical data.
- Employed a semi-supervised approach where gene selection is informed by survival time.
- SS-RPMM does not require pre-specification of the number of cancer subtypes.
Main Results:
- SS-RPMM identified cancer subtypes based on survival-associated gene subsets.
- Simulation studies demonstrated favorable performance of SS-RPMM compared to existing semi-supervised methods.
- Analysis of mesothelioma data revealed at least two distinct, survival-informative methylation profiles.
Conclusions:
- SS-RPMM is an effective tool for discovering cancer subtypes linked to survival.
- The method advances our understanding of molecular heterogeneity in cancer.
- Findings support the potential for SS-RPMM in guiding personalized cancer therapy.