Consequences of complexity within biological networks: Robustness and health, or vulnerability and disease

The phenotypes of "simple" Mendelian disorders are complex traits (1-3). As a consequence, for most genetic diseases, we are unable to accurately predict phenotype from genotype for any individual. At this time, we are limited to probabilistic estimates of individual phenotypic features, based on accumulated observations from series of individuals with a specific mutation. Therefore, if our genetic counseling is to be well-informed then it is essential that clinical and genetic data from individuals with specific mutations be accumulated and analyzed. The absence of absolute genotypic-phenotypic correlations is due to protein activity thresholds, modifier genes, and systems dynamics (1,2). In this minireview we have examined systems dynamics with a focus on network complexity and the consequences of that complexity. As we enter the genomic and proteomic era of medicine, it is essential that we increase our knowledge about the individual components of these networks, the various transcription factors, alternatively spliced and paralogous forms of individual proteins, polymorphisms including SNPs, and environmental influences on functional activity at any node. Improvements in our ability to predict phenotype from genotype, however, will require not only this detailed knowledge but also a thorough understanding of the assembly of these individual components into modules and the interrelatedness of these modules to form functioning, complex systems. The consequences of this assembly will be far greater than the sum of individual components, since, as we have seen, naturally occurring complex networks have structural features that directly influence their dynamic function. These features provide the systems with a robust tolerance to random failure of component parts, and also inform us about specific nodes of vulnerability. Understanding the vulnerabilities will be key to our ability to identify pathogenic mechanisms and intervene effectively to moderate adverse phenotypes.