User:Draperja

Introduction

The availability of high-throughput techniques opened a new epistemological era in biology. New disciplines such as genomics, functional genomics, proteomics, metabolomics and structural biology leveraged these technological advances to deepen our understanding of cellular physiology and regulation. The need to extract knowledge from these large data streams stimulated the development of computational techniques to store, organize, mine, and model the data, facilitated by the increasing availability of cheaper computer hardware and improved performance. These developments led to the emergence of the umbrella discipline of computational systems biology. The initial intent and motivation underlying the large resources and funding devoted to this major effort was to enhance human health. The human genome project (HGP), where the bulk of the work was conducted between 1990 and 2003, and which successfully led to sequencing the entire human genome, is arguably the archetypal example of the necessary combination of high-throughput techniques and computation contributing to a monumental accomplishment. The relatively modest practical output from the HGP such as advances in diagnostic testing for cancer, hematological and liver disease, and deeper understanding of comparative biology and evolution, has demonstrated that the road to translating these new disciplines into tools that would indeed contribute to diagnosing and treating human disease in a personalized fashion will be a challenging one.

Definition

One can therefore define systems medicine as the application of systems biology to the diagnosis, prevention, pathophysiologic understanding and treatment of developmental disorders, disease and recovery processes in humans. The concept of system as an assembly is fundamental: organisms are comprised of a large number of parts, or sub-systems, creating a whole which accomplishes biological functions beneficial to the integrity of the system, and admits observables that are relevant at the system level, not merely its constituent parts. In this sense, there are strong parallels between the discipline of systems engineering and systems medicine. Systems engineering is an interdisciplinary science that combines the expertise of industrial engineering, control engineering, management science towards the design, logistical execution, and maintenance of large complex projects such as submarine and airplane design, the international space station, and other large scale projects such as the internet. Importantly, systems engineering conceives of a project over its entire lifecycle. Similarly, as systems medicine strives to pursue its goals, it draws not only from rich multi-scalar data streams, but also from several quantitative fields that are not otherwise naturally aligned, such as statistics and control engineering.

Many consider that the system is not bound to individual living organisms, but also extend to communities of such organisms and how these interact. At a biological level, it is increasingly recognized that human interactions with their physical and microbiological environments has a key effect on health. An argument can be made that the goal of systems medicine is to be more global in scope and to extend beyond health of individuals. This perspective has driven large scale research efforts, such as the human microbiome project, which seeks to understand the relation between human health and the microbial flora inhabiting gut, oral cavity, upper respiratory system and skin. At the population scale and from a societal viewpoint, the interaction between healthcare delivery models and population health indicators is further pushing the envelope as to what could still be considered systems medicine. Clearly, there is an extensive need for modeling how different factors, including financial constraints, corporate and public policies, impact health at this level. Some researchers have suggested the concept of translational systems biology to describe several of the efforts, goals and promises described above and we clearly see a conceptual distinction between systems medicine and translational systems biology.

Systems medicine and personalized therapies

A fundamental goal of systems medicine is to open an evidence-based path towards personalized therapies. The concept of personalized therapy is pervasive in clinical medicine in that clinicians will implement a broad concept of treatment for specific diseases, while constantly adapting and refining treatments to the perceived circumstances of their patients. The lack of rigor in the implementation of this approach combined with a fundamental lack of scientific evidence lead to the recognition that the initial ad-hoc attempt at personalized medicine was well-intended, but prone to error, often misguided or frankly harmful. Standardizing and organizing the delivery of healthcare and building of the evidence, mostly from randomized clinical trials has started to address these concerns, but also to the realization that predictably effective personalized medicine remains a challenging long-term goal. The human genome project’s promise was to present a full description of the human genome and to leverage this knowledge towards the goal of personalized medicine. The discipline of genome medicine offers a genome-centric approach to understanding the association between the human genome, human disease, and pharmacological approaches to treatment. Genome medicine falls under the umbrella of systems medicine.

Systems medicine and models

The role of mathematical models as an integrative, formal framework that represents how constituent parts of a system are dynamically linked is more central to systems medicine than it is in more traditional systems biology, or medicine. At the very least, formalizing knowledge into a model will identify critical knowledge gaps and guide experimental efforts, including experimental design. Beyond, models represent a new vehicle through which interdisciplinary teams of quantitative, biological and clinical investigators can focus discussions and evaluate the merit of competing hypotheses prior to experimental evaluation. Models could represent the preferred method to integrate and interpret data that opens the way to personalized medicine and thus redraws the playing field for systems medicine.

Future developments

Advances in basic science and mathematics will be required for systems medicine to deliver on its promise beyond the initial results of association studies. Less than 2% of the human genome codes for proteins and the majority of proteins have regulatory roles which are not primarily involved in core cellular functions. Rather, they promote system robustness. The scientific community is getting early glimpses into the role and significance of the non-protein world and the fundamental role of epigenetics in disease. Systems medicine will likely provide a strong motivation for the development of methods and applications that will integrate this evolving knowledge in a more comprehensive theories of health and disease.

The field of pharmacokinetics has pioneered the development of model-based individualized predictions of pharmacokinetic data. Similar predictions will be much more difficult to achieve for more complex system for which experimental or clinical data are considerably sparser and for which model representations are less well known or subject to ongoing controversy. General methods that extend mixed effect modeling in standard statistical theory to non-linear dynamical systems are under development, as are model selection algorithms that allow comparing the relative merit of competing models. In several fields, such as weather prediction, the concept of model ensembles and consensus models have emerged and early approaches using similar methods to express incomplete information and other source of variations in a system in terms of parametric and structural model uncertainty. It would stand to reason that a proximal goal of this research would be to describe “similar patients” in terms of such a model ensemble, on the way to a fully probabilistic description of individual patients.

How accurate would inferences made about individual patients based on such models ensembles need to be? Scientists do not require full understanding of a system before useful applications can be realized. Gravity and electromagnetism come to mind as canonical examples of physical phenomena for which a full theory is not yet established, but applications based on incomplete knowledge have nevertheless shown extreme usefulness. How much knowledge is necessary before practical predictions can be extracted from existing models of a system is a difficult question. This applies directly to systems medicine. It is likely that current knowledge is sufficient to predict the effect of certain interventions. Several biosimulation companies have developed prediction engines for clinical trials based on in silico models of disease. Whether such contributions have actually impacted drug design on delivery is unclear at this stage, but some degree of success is almost certain within the next few years.

A roadmap for systems medicine

The construction of a roadmap for systems medicine is imperative for its development as a successful science and will require the involvement of several stakeholders, including academic faculty and trainees, the scientific dissemination industry, institutions of higher knowledge, government and other regulatory and funding entities, the biotechnology industry, and pharmaceutical companies.

Each of these promoting entities is an essential stakeholder in the systems medicine

A roadmap for systems medicine. The success of systems medicine as a discipline is contingent upon the participation of a diverse group of promoting entities, contributing enabling initiatives according to their domains of expertise and influence. The color scheme reflects the anticipated importance of their relative contribution to each initiative. (Adapted from Clermont et al.)