From functional genomics to systems biology: Our best hope to improve the treatment of complex diseases?

The phenotype of any organism is determined by three components: its genome, its environment, and some remaining random effects. Life can therefore be considered as computational process based on the information coming from the genome and the environment This computation has striking conceptual similarities to neuronal nets. Such a neuronal net, and the computation it carries out (its ‘program’) can be defined by specifying components, and their interactions, just as the genome of an organism specifies the components of the molecular machinery (genes, transcripts, proteins), and, specified indirectly by the sequence and structure of these components, their interactions. The shaping of the genome in evolution, in this sense, corresponds to the redesign/reprogramming of a neuronal net. In our view this analogy has major repercussions on what we can hope to achieve by studying disease processes. Since a neuronal net cannot be ‘understood’ in the classical sense, we similarly might never be able to ‘understand’ many of the complex disease processes. We are however able to predict the response of a neuronal net to specific disturbances, if we are able to model the computational process, essentially duplicating the neuronal net in all its complexity in the computer. Using such models we can, for example, analyse the response of a neuronal net, which is not directly accessible (e.g. on a mars robot) to specific disturbances (e.g. breakdown of specific components) and explore ways to correct such defects, by changing the pattern of interconnection between the components, which continue to work. Similarly we might never be able to ‘understand’ many of the common, complex diseases like cancer, heart disease or neuronal diseases. We should however be able to identify the components acting in these processes, and to get at least a rough picture of their interactions both in the healthy organisms, and the disease state, using functional genomics, and to use this information to model the complex network underlying the disease, and the effect, the disease has on it, using systems biology approaches. Such a model could conceivably allow us to achieve a real ‘individualised’ medicine. It should for example be possible to derive models of many disease processes, and to adapt the model to each individual patient, based on many different sources of information (genetic/genomic, expression pattern at the RNA and protein level, analysis of metabolic profiles etc.). Such a model could for example, be used to predict the response of each individual patient to each possible therapy, and therefore to select the therapy, expected to have the highest effectiveness, and minimal side effects. Similarly the availability of exact, predictive models of the disease process for important diseases could have a major effect on disease prevention, improve diagnosis, and revolutionise the development of new drug. We consider this development of a medical systems biology an essential new step in increasing the health of the population, in improving medical treatment, and indirectly also in lowering the steadily increasing health costs of an ageing population. It is a logical next step after the molecular revolution, providing us with information on many of the basic mechanisms acting in diseases and the genomic revolution, allowing us to identify and characterise most or all of the components of the complex networks at the root of many common, complex diseases. Only the next step, medical systems biology, will however provide the predictive capability, which we urgently need to solve many of the real world problems we are facing, and will continue to face for the foreseeable future.