Batch Effect Correction to Enhance Utility of a Publicly Available ALS Proteomics Dataset

Batch effect was identified in the Answer ALS 1-2 proteomics dataset (508 samples × 3,643 proteins), and BDC-ComBat reduced it globally while preserving key neuronal PC-contributing proteins.

• Why it matters: Public ALS multi-omic resources can be hard to reproduce when batch effects obscure true biological signal. This work targets that gap for the Answer ALS proteomics dataset.
• What they did: The study compared Answer ALS batch differential controls with ComBat, testing ComBat-BDC versus BDC-ComBat across the dataset and examining outliers MOS and H4 and PC-contributing proteins.
• The result: BDC-ComBat was the only approach that reduced batch effect across the entire dataset and produced a corrected version with 354 samples × 3,641 proteins, recapitulating neuronal proteins including AINX, NFL, and NFM.

AI summary of the abstract · may contain errors · not medical advice · read the paper.