PT - JOURNAL ARTICLE AU - Adithya Sagar AU - Wei Dai AU - Mason Minot AU - Jeffrey D Varner TI - Reduced order modeling and analysis of the human complement system AID - 10.1101/059386 DP - 2016 Jan 01 TA - bioRxiv PG - 059386 4099 - http://biorxiv.org/content/early/2016/06/16/059386.short 4100 - http://biorxiv.org/content/early/2016/06/16/059386.full AB - Complement is an important pathway in innate immunity, inflammation, and many disease processes. However, despite its importance, there have been few validated mathematical models of complement activation. In this study, we developed an ensemble of experimentally validated reduced order complement models. We combined ordinary differential equations with logical rules to produce a compact yet predictive complement model. The model, which described the lectin and alternative pathways, was an order of magnitude smaller than comparable models in the literature. We estimated an ensemble of model parameters from in vitro dynamic measurements of the C3a and C5a complement proteins. Subsequently, we validated the model on unseen C3a and C5a measurements not used for model training. Despite its small size, the model was surprisingly predictive. Global sensitivity and robustness analysis suggested complement was robust to any single therapeutic intervention. Only the knockdown of both C3 and C5 consistently reduced C3a and C5a formation from all pathways. Taken together, we developed a reduced order complement model that was computationally inexpensive, and could easily be incorporated into pre-existing or new pharmacokinetic models of immune system function. The model described experimental data, and predicted the need for multiple points of therapeutic intervention to fully disrupt complement activation.