A cellular stress response and an adverse clinical outcome are different observations. A 2017 toxicogenomic study examined how dose and time shaped inferred gene-regulatory networks in chemically exposed human hepatocytes.
The researchers analyzed publicly available expression data using Dose–Time Network Identification. They compared groups of compounds associated with carcinogenicity, drug-induced liver injury, and a comparator category labeled noncarcinogenic and non-DILI. Four stress pathways were examined, including NF-kappa B, NRF2, TP53, and endoplasmic-reticulum stress.
The comparator group also produced an inferred disturbance in the NF-kappa B network. The other chemical groups showed broader pathway effects and different inferred interactions. Thus, the presence of a stress-pathway response alone did not reproduce the distinction between the study’s outcome categories.
The networks were inferred from expression patterns across experimental conditions. An inferred relationship between genes is a mechanistic hypothesis requiring evidence appropriate to that relationship. The study’s chemical categories also describe its analytical grouping, with limited grounds for treating a comparator as harmless in every setting. For forensic toxicologists reviewing a molecular injury claim, identify the measured expression change, tested concentrations, timing, model assumptions, and separate evidence of tissue injury. A signal may reflect adaptation, adverse processes, or overlapping responses. Connecting it to an individual’s disease requires more than detecting a pathway change in an experimental dataset.
Read the full scientific paper. DOI 10.3389/fgene.2017.00142.