Abstract
The discovery of molecular prognostic cancer biomarkers is still a major scientific challenge. Some methodologies have been proposed to generate novel model biomarkers for clinical outcome using gene expression as predictors but involve some drawbacks. For example, (i) they heavily depend on a rank of the initial univariate relation to survival times, (ii) are unable to generate compact multivariate predictors, (iii) are based on survival models other than Cox, or (iv) use aggregation and transformations of expression values instead of the gene expression directly. These issues complicate the evaluation of biomarkers in clinical trials, its implementation in medical practice and obscures its biological association with cancer.
We propose a particle swarm optimization search engine coupled to multivariate Cox survival model fitting, constraining the number of genes while minimizing for deviance residuals to identify prognostic biomarkers cancer. By evaluating the concordance index, Log-rank, correlation, the integrated discrimination improvement per feature and the number of variables significantly associated to survival times, we show that many compact and highly predictive models can be found for six cancer datasets and a simulated cohort. We also show that our algorithm generates a competitive population of multivariate models with a wide variety of gene combinations, including genes that could not be found by a univariate methodology. In comparisons with other methods such as LASSO, Ridge, and Elastic Net, our algorithm shows similar or better results.
We conclude that our algorithm generates highly predictive and compact models for clinical outcomes with a unique gene content, and a superior or comparable prediction to other current feature selection methods. R and Java code are available in Supplementary Information and http://bioinformatica.mty.itesm.mx/?q=coxswarm.
Keywords: Clinical outcome, microarrays, gene expression, feature selection, biomarkers.
Current Bioinformatics
Title:Efficient Gene Selection for Cancer Prognostic Biomarkers Using Swarm Optimization and Survival Analysis
Volume: 11 Issue: 3
Author(s): Raul Aguirre-Gamboa, Emmanuel Martinez-Ledesma, Hugo Gomez-Rueda, Rebeca Palacios, Isabel Fuentes-Hernandez, Emilio Sánchez-Canales, Rafael Chacolla-Huaringa, Servando Cardona-Huerta, Luis Villela, Sean-Patrick Scott, Jose Tamez-Pena and Victor Trevino
Affiliation:
Keywords: Clinical outcome, microarrays, gene expression, feature selection, biomarkers.
Abstract: The discovery of molecular prognostic cancer biomarkers is still a major scientific challenge. Some methodologies have been proposed to generate novel model biomarkers for clinical outcome using gene expression as predictors but involve some drawbacks. For example, (i) they heavily depend on a rank of the initial univariate relation to survival times, (ii) are unable to generate compact multivariate predictors, (iii) are based on survival models other than Cox, or (iv) use aggregation and transformations of expression values instead of the gene expression directly. These issues complicate the evaluation of biomarkers in clinical trials, its implementation in medical practice and obscures its biological association with cancer.
We propose a particle swarm optimization search engine coupled to multivariate Cox survival model fitting, constraining the number of genes while minimizing for deviance residuals to identify prognostic biomarkers cancer. By evaluating the concordance index, Log-rank, correlation, the integrated discrimination improvement per feature and the number of variables significantly associated to survival times, we show that many compact and highly predictive models can be found for six cancer datasets and a simulated cohort. We also show that our algorithm generates a competitive population of multivariate models with a wide variety of gene combinations, including genes that could not be found by a univariate methodology. In comparisons with other methods such as LASSO, Ridge, and Elastic Net, our algorithm shows similar or better results.
We conclude that our algorithm generates highly predictive and compact models for clinical outcomes with a unique gene content, and a superior or comparable prediction to other current feature selection methods. R and Java code are available in Supplementary Information and http://bioinformatica.mty.itesm.mx/?q=coxswarm.
Export Options
About this article
Cite this article as:
Aguirre-Gamboa Raul, Martinez-Ledesma Emmanuel, Gomez-Rueda Hugo, Palacios Rebeca, Fuentes-Hernandez Isabel, Sánchez-Canales Emilio, Chacolla-Huaringa Rafael, Cardona-Huerta Servando, Villela Luis, Scott Sean-Patrick, Tamez-Pena Jose and Trevino Victor, Efficient Gene Selection for Cancer Prognostic Biomarkers Using Swarm Optimization and Survival Analysis, Current Bioinformatics 2016; 11 (3) . https://dx.doi.org/10.2174/1574893611999160610125628
DOI https://dx.doi.org/10.2174/1574893611999160610125628 |
Print ISSN 1574-8936 |
Publisher Name Bentham Science Publisher |
Online ISSN 2212-392X |
![](/images/wayfinder.jpg)
- Author Guidelines
- Bentham Author Support Services (BASS)
- Graphical Abstracts
- Fabricating and Stating False Information
- Research Misconduct
- Post Publication Discussions and Corrections
- Publishing Ethics and Rectitude
- Increase Visibility of Your Article
- Archiving Policies
- Peer Review Workflow
- Order Your Article Before Print
- Promote Your Article
- Manuscript Transfer Facility
- Editorial Policies
- Allegations from Whistleblowers
Related Articles
-
New Pathways for Reactive Oxygen Species Generation in Inflammation and Potential Novel Pharmacological Targets
Current Pharmaceutical Design Formulation and Characterization of Rutin Loaded Chitosan-alginate Nanoparticles: Antidiabetic and Cytotoxicity Studies
Current Drug Delivery Graphical Abstracts
Letters in Drug Design & Discovery Dysregulation of Lysyl Oxidases Expression in Diabetic Nephropathy and Renal Cell Carcinoma
Current Drug Targets Titanocene Y and Vanadocene Y: Platinum Resistance-Breaking Cytotoxic and DNA-Targeting Anticancer Drug Candidates
Letters in Drug Design & Discovery Dissecting Bortezomib: Development, Application, Adverse Effects and Future Direction
Current Pharmaceutical Design Telomerase Modulation in Therapeutic Approach
Current Pharmaceutical Design Iron Chelating Strategies in Systemic Metal Overload, Neurodegeneration and Cancer
Current Medicinal Chemistry Deregulation of the Akt Pathway in Human Cancer
Current Cancer Drug Targets Resveratrol-Mediated Reversal of Tumor Multi-Drug Resistance
Current Drug Metabolism Innovations and Improvements in Pharmacokinetic Models Based on Physiology
Current Drug Delivery Strategies for Target-Specific Contrast Agents for Magnetic Resonance Imaging
Current Molecular Imaging (Discontinued) The Recombinant T Cell Receptor Strategy: Insights into Structure and Function of Recombinant Immunoreceptors on the Way Towards an Optimal Receptor Design for Cellular Immunotherapy
Current Gene Therapy Bioprecursor Prodrugs: Molecular Modification of the Active Principle
Mini-Reviews in Medicinal Chemistry The Metastatic Process: Methodological Advances and Pharmacological Challenges
Current Medicinal Chemistry HE4 is a Beneficial Biomarker in Endometrial Adenocarcinoma
Current Women`s Health Reviews Notch Signalling Pathways and Their Importance in the Treatment of Cancers
Current Drug Targets Neutrophil Gelatinase-Associated Lipocalin (NGAL) as a New Biomarker for Non – Acute Kidney Injury (AKI) Diseases
Inflammation & Allergy - Drug Targets (Discontinued) Shedding Light on Serum Vitamin D Concentrations and the Risk of Rarer Cancers
Anti-Cancer Agents in Medicinal Chemistry Antisense Oligodeoxynucleotide Therapy for Prostate Cancer Targeting Antiapoptotic Genes Involved in the Mechanism Mediating Progression to Androgen Independence
Current Signal Transduction Therapy