Abstract
The interaction between a compound and a biological system changes across the concentration gradient: it is selective and target-mediated at low concentrations and non-specific and stress-dominated at high concentrations. However, bioactivity screening favours concentrations high enough to produce clear, statistically robust effects, which preferentially captures the stress-dominated regime and overlooks the range in which regulated, adaptive responses occur. This review examines how sparse concentration spacing, monotonic analytical models and effect-size-driven selection interact to bias discovery towards high-concentration bioactivity. The biology of the response itself may compound the problem: above the baseline toxicity threshold, structurally unrelated compounds converge on shared damage pathways, producing directionally consistent outcomes that are statistically convenient to detect. In contrast, responses below the threshold preserve specificity, but depend more strongly on biological state and experimental context. High-concentration effects also frequently fail to translate, due to constraints imposed by matrix interactions, cost, regulatory limits, and loss of specificity that screening does not anticipate. Hormesis, when viewed as an emergent property of regulatory networks rather than an inherent property of the compound, provides a framework for these qualitative differences. As an alternative organising principle, the concept of a functional window is proposed: a concentration range defined not by maximal effect, but by reproducibility, biological coherence, the absence of sustained stress signatures, temporal robustness and operational compatibility. The criteria for identifying and validating functional windows are presented, offering a structured approach to recovering low-concentration bioactivity relevant to the discovery and development of applied biomolecules.