BERT

BERT provides efficient batch-effect adjustment of data with missing values by ordering all batch effect correction into a tree of pairwise computations with parallelization over sub-trees.

Key Features:

• Batch-Effect Reduction Trees: BERT orders all batch effect correction into a tree of pairwise computations.
• Parallelization: BERT allows parallelization over sub-trees for computational efficiency.
• Missing Value Handling: BERT adjusts data containing missing values without requiring complete data matrices.

Scientific Applications:

• Large-Scale Omic Data Integration: BERT enables high performance data integration for large-scale analyses of incomplete omic profiles.
• Batch Effect Correction: BERT removes unwanted technical variation from heterogeneous datasets compiled from multiple batches or sources.

Methodology:

BERT constructs a tree of pairwise batch effect computations and applies parallelization over sub-trees to achieve efficient correction of data with missing values.