Abstract

Mathematical and computational modeling enables in silico testing of (for example) hypotheses, experimental design, and interventional strategies. However, building, sharing, and applying complex models require technical skills and software development knowledge that is rarely accessible outside computational groups. This barrier limits broader adoption, reproducibility, and collaborative model development which, unfortunately, slows the pace of discovery. In their recent paper, Johnson and colleagues address this challenge by introducing a human-interpretable grammar that encodes multicellular systems biology models as human-readable statements following a fixed structure (e.g., "cell type, signal, response, behavior"). This framework allows investigators to speak mathematical models into existence by composing biological hypotheses in plain-text sentences that are then translated into executable agent-based models. The authors show, through a set of cancer-relevant examples, how this grammar bridges the gap between biological reasoning and mathematical formalism, thereby providing a language that can be shared between experimentalists, computational scientists, and clinicians. Practically, this means that models can be composed, modified, and reproduced without requiring programming expertise. This democratization could accelerate discovery, broaden participation in computational oncology, and facilitate the translation of modeling insights into experimental and clinical research. Thus, Johnson and colleagues have lowered the barrier for model construction and application.