Confidence calibration across RNA-seq modalities
#1
by Shicheng - opened
Harmonizing 1,037,214 human RNA-seq runs across 49,246 studies—and retaining field-level confidence scores plus ontology mappings—addresses a major obstacle in public transcriptomic reuse. How have you assessed whether confidence thresholds control annotation errors consistently across bulk, single-cell, and spatial studies when constructing disease-focused discovery cohorts?