ABDS

ABDS analyzes biologically diverse samples through missing data imputation, signature gene detection, and differential pattern visualization.

Key Features:

• Mechanism-integrated group-wise pre-imputation: A pre-imputation scheme retains informative missingness associated with signature genes.
• Cosine-based one-sample test: An extended test detects group-silenced signature genes.
• Unified heatmap: A heatmap design displays multiple sample groups simultaneously.
• Mechanistic imputation: Tools perform imputation accounting for mixed missing mechanisms.
• Group-wise testing: Statistical testing compares multiple phenotypic sample groups in parallel.
• Heatmap visualization: Visualization tools display differential patterns across sample groups.

Scientific Applications:

• Missing value imputation: Handles high missing rates with mixed missing mechanisms in biologically diverse samples.
• Signature gene detection: Identifies informative molecular features that define different phenotypic sample groups, including group-silenced signature genes.
• Differential pattern visualization: Displays multiple sample groups for comparative analysis of molecular signals.
• Biomedical data analysis: Enables accurate detection of interpretable molecular signals among phenotypically diverse sample groups.

Methodology:

ABDS implements a mechanism-integrated group-wise pre-imputation scheme, extends a cosine-based one-sample test for group-silenced signature gene detection, and designs a unified heatmap for multiple sample group display. The methods are implemented as an R package.