Sstack

Sstack integrates heterogeneous genomic feature sets to enable predictive modeling when subsets of genomic features are missing across samples.

Key Features:

• Heterogeneous Data Integration: Handles scenarios where different subsets of genomic features are missing in parts of a dataset, enabling analysis of incomplete datasets without compromising model integrity.
• Sequential Addition of Samples and Features: Supports sequential addition of samples and features to accommodate incremental data collection and dynamic information integration.
• Improved Prediction Accuracy: Stacks genomic datasets to improve predictive accuracy, as demonstrated in drug sensitivity prediction using the Cancer Cell Line Encyclopedia (CCLE).
• R Package Implementation: Implemented as an R package for use within R-based computational workflows.

Scientific Applications:

• Drug Sensitivity Prediction: Enhances accuracy of drug response predictions, demonstrated using CCLE data to inform precision medicine analyses.
• Genomic Data Analysis: Enables analysis of complex genomic datasets with missing features to support biological insight and downstream modeling.

Methodology:

Sstack stacks different genomic feature sets to create a unified model that compensates for missing data and supports the sequential addition of samples and features.