Corrado Priami è Presidente e Amministratore delegato di COSBI [centre for COmputational and Systems BIology] dal 2005 e professore ordinario presso il dipartimento di Matematica dell’Università di Trento. Melissa J. Morine, dottore di ricerca in bioinformatica e nutrigenomica, collabora con COSBI dal 2011.
The convergence between computer science and biology has occurred in successive waves, involving increasingly deeper concepts of computing. Since its early days, computer science has taken inspiration from nature, with the works of Turing, Von Neumann and Minsky. These milestones led to extraordinary results, some of which recall biology even in their names: cellular automata, neural networks, genetic algorithms. The current situation makes computer science a suitable candidate for becoming a foundation for systems biology with the same importance as mathematics, chemistry and physics. The same applies to more applied systems biology domains related to the healthcare sector such as systems nutrition (which studies the impact of nutrients on cellular machinery) or pharmacology.
Biology is experiencing tremendous growth in the capacity to measure biological molecules and processes with unprecedented depth and precision. The resulting inundation of data presents as many challenges as it does opportunities, such that organization and analysis of these data requires specialized training and dedicated research. Statistics and computer science are now fundamental to biological data analysis, as it has become commonplace to analyze terabytes of data that comprehensively characterize complex biological systems. Another defining feature of modern biological research is a growing interest in interpreting living systems as dynamic information manipulators, and as such is moving toward systems biology.
Computing and biology have been converging ever more closely for the past two decades, but with a vision of computing as a resource for biology that has propelled bioinformatics. Bioinformatics addresses structural and static aspects of biology and has produced databases, pattern manipulation and comparison, search tools and data mining techniques. Computational approaches in biology are now moving towards systems biology (see Systems biology box). This poses both challenges and opportunities to describe the step-by-step mechanistic behavior that underlies complex phenotypes.
— Systems biology
There is no universal agreement on a definition of systems biology. We consider systems biology a transition from
- qualitative biology to a quantitative science;
- reductionism to system level understanding of biological phenomena;
- structural, static descriptions to functional, dynamic properties; and
- descriptive biology to mechanistic/causal biology.
Some of the main aspects of interest in the transition are causality between events, temporal ordering of interactions and spatial distribution of components within the reference volume of reactions.
The current approach in biological data collection is to take snapshots of biological systems and subsequently try to model the variation of measures in the snapshots through equations. Such snapshots may represent a wide variety of systems and states, from a simple in vitro cellular model at multiple time points, to a series of human tissue samples across a range of disease states. These snapshots contain a great deal of information of the state of a given system, but can be increasingly complicated to analyze and interpret, and thus require dedicated statistical and bioinformatic strategies. Naturally, system snapshots do not directly capture the dynamics that carry the system from one state to the next. Algorithms describe the steps between system states in a causal continuum of actions that make the measures change, thus providing a dynamic view of the system under question. Through algorithms and the (programming) languages used to specify them, we can recover temporal, spatial and causal information on the modeled systems by using well-established computing techniques that deal with program analysis, composition and verification, integrated software development environments and debugging tools as well as algorithm animation.
An algorithm is a finite list of well-defined instructions for allowing an executor to perform a task without ambiguity. Programming languages are used to express algorithms when the executor is a digital computer.
Algorithms need a syntax to be described and a semantics to associate them with their intended meaning so that an executor can precisely perform the steps needed to implement the algorithms with no ambiguity.
Algorithms are quantitative when the selection mechanism of the next step is determined according to probabilistic/temporal distributions associated with either the rules or the components of the system modeled.
Hereafter we call algorithmic systems biology the specification of biological models through algorithms to describe their step-by-step dynamics.
The main difference between algorithmic systems biology and other techniques used to model biological systems stems from the intrinsic difference of algorithms (operational descriptions) and equations (denotational descriptions). An equation might be an elegant way of describing the result of the execution of an algorithm. Furthermore, equations specify dynamic processes by abstracting the steps performed by the executor, thus hiding from the user the causal, spatial and temporal relationships between the elementary steps. Equations describe the variation of variables (usually concentrations of species) from one state to another of a system, while algorithms highlight why and how a system moves from one state to another one.
Algorithms force modelers/biologists to think about the mechanisms governing the behavior of the system under question. Algorithms can serve to coherently extract general biological principles that drive the data produced in systems biology, and are a practical tool for expressing and favoring computational thinking. Therefore, they are also a conceptual tool that helps to understand fundamental biological principles. Statistical modeling and analysis are essential in this task, for extracting the relevant biological signal from high-throughput data that may initially be too complex to be modeled by algorithms.
The notion of simulation needs some consideration. Algorithmic simulations are executable on computers and rely on deep computing theories, while mathematical simulations are solved with the support of computer programs (hence, computing here is just a service). Execution of algorithms exhibits the emergent behavior produced at system level through the set of local interactions between components without the need of specifying it from the beginning. The complex interaction of the concentrations of the species, the sensitivity of their interactions expressed through stochastic parameters, the localization of the components in a three dimensional hierarchical space, and hence the dynamic evolution of a system can all be modeled through computational simulation.
A simulation is the process of model solution/execution to reproduce an approximation of the dynamics of a system through a solver/executor (usually a digital computer).
The outcome of a simulation is an approximation of the time behavior of a system.
Overall, the integration of many different disciplines like mathematics and statistics to extract knowledge from data, network theory to identify functional biological subnetworks from data, computing and mathematics to describe the dynamics of systems and simulate their temporal behavior, and biochemistry, biology and medicine to provide the background knowledge for interpretation of modeling results are all mandatory to succeed in systems biology.
Courtesy by ICP | www.icpress.co.uk