OCSVM

OCSVM identifies disease-associated genes from gene expression data using one-class support vector machine classification to prioritize known disease-causing genes, with demonstrated application to acute myeloid leukemia (AML) datasets.

Key Features:

• One-Class Classification Focus: Uses one-class support vector machines to model only known disease genes and detect anomalous (disease-associated) patterns without requiring labeled non-disease genes.
• Enhanced Sensitivity and Precision: Prioritization of known disease genes improves accuracy metrics, with reported gains in precision, recall, and F-measure on AML gene expression datasets.
• Benchmarking and Validation: Performance was evaluated using a benchmark dataset derived from gene expression profiles associated with acute myeloid leukemia to assess robustness.

Scientific Applications:

• Gene Prioritization in Genomics: Streamlines identification of genes implicated in diseases such as acute myeloid leukemia from gene expression data.
• Therapeutic Target Nomination: Narrows candidate genes for therapeutic intervention by improving precision and recall in disease-gene detection.
• AML Expression Analysis: Applicable to analysis and interpretation of acute myeloid leukemia gene expression datasets for disease-associated signal detection.

Methodology:

Constructs a model using one-class support vector machines applied to gene expression profiles to distinguish normal versus anomalous (disease-associated) patterns without requiring negative training samples.