Oncogenes and pathway identification using filter-based approaches between various carcinoma types in lung

Area of Science:

• Oncology
• Bioinformatics
• Computational Biology

Background:

• Lung cancer remains a leading cause of cancer mortality globally.
• Identifying cancer-associated genes and pathways is crucial for effective cancer prevention and treatment.
• Existing methods for lung cancer subtype identification require systematic improvement.

Purpose of the Study:

• To systematically identify key genes and pathways implicated in lung cancer.
• To develop and evaluate advanced computational methods for lung cancer subtype classification.
• To enhance the accuracy and efficiency of lung cancer diagnosis through gene prioritization.

Main Methods:

• Pathway analysis utilizing Hypergeometric Distribution (HGD) to identify significantly overrepresented reaction sets.
• Application of feature selection techniques including Particle Swarm Optimization (PSO), Information Gain (IG), and Biomarker Identifier (BMI) for lung cancer identification.
• Development of a novel method for determining Biomarker Identifier (BMI) thresholds to prioritize genes.

Main Results:

• Identification of key gene sets present across multiple significant pathways.
• Effective feature simplification achieved through the proposed computational approach.
• Demonstrated higher classification accuracy for lung cancer subtypes compared to existing literature methods.

Conclusions:

• The integrated approach effectively identifies critical genes and pathways in lung cancer.
• The developed methods offer improved feature selection and gene prioritization for lung cancer diagnostics.
• This systematic strategy enhances classification accuracy, paving the way for more precise lung cancer management.