Auditing Retrieval-Augmented LLM Hypotheses for Longitudinal Cell Painting Morphology

2026/06/30 by Gilchan Park, G Y Park, Guang Zhao +2
Biochemistry, Genetics and Molecular Biology · Computer Science · Engineering · #3D Printing in Biomedical Research #Audit #Biomedical Text Mining and Ontologies #Cell Image Analysis Techniques #Cell survival #Identifier #Morphology (biology) #Systems biology #cs.AI #cs.CL #q-bio.QM

paper · pdf · doi:10.1145/3807503.3819448

openalex publication_date 2026/06/30 · openalex created_date 2026/07/25 · openalex updated_date 2026/08/01

Abstract

High-content morphological profiling (Cell Painting) yields sensitive, high-dimensional signatures of cellular state, but translating longitudinal morphology trajectories into interpretable biology remains difficult—especially for weak, chronic perturbations such as low-dose-rate ionizing radiation. Large language models (LLMs) can synthesize heterogeneous evidence into narratives, yet scientific use requires quantitative auditing: fluent outputs may be ungrounded or inconsistent with measured morphology.

Citations

• On the Dangers of Stochastic Parrots

Related

• Causal integration of chemical structures improves representations of microscopy images for morphological profiling
• CellCLIP -- Learning Perturbation Effects in Cell Painting via Text-Guided Contrastive Learning
• A Decade in a Systematic Review: The Evolution and Impact of Cell Painting
• ViTally Consistent: Scaling Biological Representation Learning for Cell Microscopy
• SciHorizon-GENE: Benchmarking LLM for Life Sciences Inference from Gene Knowledge to Functional Understanding