LiquidTransformer — the first liquid biopsy-native AI platform.
Designed to capture the faint signals from cell-free DNA fragments in chronic disease, our patented transformer-based model uses the attention mechanism to determine fragment significance at a single-molecule level. Unlike traditional liquid biopsy methods, LiquidTransformer captures the molecular signal in blood that reflects the molecular state of the tissue. Validated in our MASH Atlas — a paired plasma–tissue multiomics study.
AASLD 2025 oral podium presentation · manuscript in submission