Omics approaches to individual variation: modeling networks and the virtual patient

Abstract

Every human is unique. We differ in our genomes, environment, behavior, disease history, and past and current medical treatment—a complex catalog of differences that often leads to variations in the way each of us responds to a particular therapy. We argue here that true personalization of drug therapies will rely on “virtual patient” models based on a detailed characterization of the individual patient by molecular, imaging, and sensor techniques. The models will be based, wherever possible, on the molecular mechanisms of disease processes and drug action but can also expand to hybrid models including statistics/machine learning/artificial intelligence-based elements trained on available data to address therapeutic areas or therapies for which insufficient information on mechanisms is available. Depending on the disease, its mechanisms, and the therapy, virtual patient models can be implemented at a fairly high level of abstraction, with molecular models representing cells, cell types, or organs relevant to the clinical question, interacting not only with each other but also the environment. In the future, “virtual patient/insilico self” models may not only become a central element of our health care system, reducing otherwise unavoidable mistakes and unnecessary costs, but also act as “guardian angels” accompanying us through life to protect us against dangers and to help us to deal intelligently with our own health and wellness.

Mathematical modeling of antihypertensive therapy

“…People differ in their genomes, environments, behaviours, and disease histories—all of these differences lead to variations in response to a particular medical treatment. Thus, true personalization of drug therapies should rely on “virtual patient” models implemented at the level of abstraction required for a specific pathology ( Lehrach, 2016 ). Here we present a technology for constructing a virtual patient and an algorithm for generating a virtual population with varying values of the necessary parameters.…”

Thoroughly Calibrated Modular Agent-Based Model of the Human Cardiovascular and Renal Systems for Blood Pressure Regulation in Health and Disease

“…Step 1 : Creation of a mathematical model of human physiology with a level of detail sufficient to study the issues of arterial hypertension . True personalization of drug therapies should rely on a virtual patient, the digital twin of a real individual, which is formed and accumulated throughout his life as a result of interaction with the health care system ( Lehrach, 2016 ). We suppose that it is unrealistic to build a virtual patient for all occasions now.…”