AI Insight
Researchers developed a framework that uses large language models (LLMs) to interpret complex cellular imaging data from cells exposed to low-dose radiation over 9 weeks. The system combines morphological changes with scientific literature and other evidence sources to generate testable biological hypotheses, and introduces two quantitative methods to verify the accuracy of LLM-generated interpretations. Applied to cells exposed to various radiation dose rates, the framework identified potential adaptive responses involving metabolic changes and cellular stress at lower radiation doses.
Why it matters
This work addresses a critical challenge in using AI for scientific discovery by creating verifiable methods to audit LLM outputs, making them more trustworthy for biological research. The framework could accelerate hypothesis generation from high-dimensional imaging data and improve understanding of chronic low-dose radiation effects, which has implications for radiation safety standards and space exploration.
Understand the Science
arXiv:2607.19415v1 Announce Type: new
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 biological narratives, yet their scientific use requires quantitative auditing. We present an evaluation-first, retrieval-augmented interpretation framework for longitudinal Cell Painting morphology, applied to a 9-week RPE-1 time course across five dose rates (0.003–6.0 mGy/hr). Week-matched treated-control morphology deltas are combined with retrieved perturbation neighbors, pathway context, and literature evidence through stable evidence identifiers, enabling an LLM to generate structured, evidence-linked hypotheses that are hierarchically summarized while preserving provenance. We introduce two quantitative auditing tests: V1 citation validity, which verifies that cited evidence identifiers exist in the prompt, and V2 proxy-based morphology compatibility, which evaluates consistency between predicted biological processes and the most altered morphology features. In our experiments, V1 detected no invalid evidence references, while V2 showed meaningful morphology compatibility that increased with perturbation strength and was positively associated with an independent morphology drift summary. The framework produces auditable, falsifiable biological hypotheses, including an adaptive phenotype involving metabolic reprogramming and proteostatic stress at lower dose rates (0.003–0.3 mGy/hr). Current limitations include proxy-based evaluation and the lack of ground-truth mechanism labels.
Source: Auditing Retrieval-Augmented LLM Hypotheses for Longitudinal Cell Painting Morphology