mAPKL

mAPKL performs hybrid gene selection and downstream analysis of microarray gene expression data to identify exemplar genes for classification, network analysis, and pathway interpretation.

Key Features:

• mAP-KL hybrid gene selection: Implements the mAP-KL method to select small sets of exemplar genes from microarray data while preserving classification accuracy.
• Multiple hypothesis testing integration: Incorporates multiple hypothesis testing into the gene selection workflow.
• Affinity Propagation (AP) clustering: Uses the affinity propagation clustering algorithm to group genes and identify exemplars.
• Krzanowski & Lai cluster quality index: Applies the Krzanowski & Lai index to assess cluster quality within the AP-clustering step.
• Data import: Supports import of raw microarray data for downstream analysis.
• Data sampling: Allows sampling of data according to user-defined proportions for dataset management.
• Preprocessing (normalization and transformation): Provides normalization and transformation options for microarray preprocessing.
• Support Vector Machine (SVM) classification: Incorporates SVM for classification of samples based on selected genes.
• Performance evaluation metrics: Includes evaluation metrics to assess classification accuracy and reliability.
• Network analysis: Performs network analysis on selected exemplar genes including degree centrality, closeness, betweenness, and clustering coefficient, and constructs edge list tables.
• Gene annotation: Provides gene annotation analysis for functional interpretation of selected genes.
• Pathway analysis: Supports pathway analysis to explore biological pathways associated with exemplar genes.
• Automated reporting: Generates automated analysis reports summarizing selection, classification, and network results.

Scientific Applications:

• Microarray gene expression profiling: Identification of small informative gene sets for analysis of microarray datasets.
• Biomarker and exemplar selection: Selection of exemplar genes for use in classifier development and validation.
• Cancer research: Application to cancer-related microarray studies for gene selection, classification, and pathway interpretation.
• Genetic disorder analysis: Use in studies of genetic disorders where reduced gene sets and network context aid interpretation.
• Network and pathway interpretation: Analysis of gene interaction networks and associated pathways for functional insights.

Methodology:

Methods explicitly stated include data import, data sampling, normalization and transformation, multiple hypothesis testing, the mAP-KL hybrid method combining multiple hypothesis testing with Affinity Propagation clustering using the Krzanowski & Lai cluster quality index to select exemplars, Support Vector Machine classification with performance evaluation metrics, network analysis computing degree centrality, closeness, betweenness and clustering coefficient plus edge list construction, gene annotation and pathway analysis, and automated report generation.