Introduction

The information delivered by genome sequencing, transcriptome, proteome and metabolome analysis, including protein/protein and protein/nucleic acid interaction maps, is so extensive and complex, that its runs the danger of undergoing the same fate as much of the scientific information gathered in the last century: to be accumulated but not applied. The activity of putting the information back together in such a way that a model of a functioning organism is obtained, will constitute a tremendous ordering force. The data will be linked organically. The definition of molecular blueprints that describe cellular events, will structure the experimental findings and foster quantitative predictive ability.

The budding yeast is the organism of choice for this ambitious goal given its relatively small genome, extensive genetic information, well-known biochemistry and proven first steps of turning it into a Silicon Cell.

It is through their dynamic interactions that macromolecules gives rise to the functioning living cell and organism. These dynamic interactions cannot be derived from their primary sequence alone; additional kinetic, expression-level, and topological information is required. Only for two organisms anywhere near sufficient such information is available to beging making replica in silico, i.e. the bacterium Escherichia coli and the yeast S. cervisiae. These two organisms also have the advantage of optimal experimental accessibility, unicellularity and comparative simplicity. Of these two organisms the latter is much closer to the ultimate aim of making Silicon cells, i.e. human cells. The most important intracellular processes such as metabolism, signal transduction, energetics, cell cycling, senescence, chromatin regulated gene-expression, drug metabolism, and apoptosis are conserved from yeast to man, making this organism a suitable model system for many aspects of human cell function. Consequently, the identification of the molecular blueprint of all yeast cellular functions would be of extreme interest both for drug discovery and for biotechnology more in general.

What is new in the Silicon cell concept?

S. cerevisiae has been the among the first two organisms for which a part existed in silico [Teusink et al, Bakker et al, Reuss et al]. All the enzymes of its major catabolic pathway were characterized kinetically under the single experimental condition that is relevant for the inside of the cell. The kinetic characteristics were then translated into computer code and integrated. This Silicon replica of glycolysis in yeast was then compared with the real cells.

These parts of Silicon cells are already available for in silico experimentation on a web site of this consortium: http://www.jjj.bio.vu.nl/

There have been many computer models of parts of living cells in the literature. Why then were these more recent models the first Silicon Cell models? They were the first because they calculated exclusively on the basis of the required experimental kinetic data.

In the calculations there was no fitting. The making of Silicon Cells therefore does not correspond to modeling (when modeling one plays with parameter values and structures until one gets behavior corresponding to the observations) but to calculation. In this sense, the making of Silicon cells is Bioinformatics, i.e. calculating understanding from nothing but data.

Recently, a first-stage blueprint of cell-cycle events was developed, much along the same lines. The silicon cell cycle behaved quite similarly to experimental cells in terms of population dynamics in steady and transitory states (1). Alteration of relevant metabolic parameters by environmental and genetic means (input/output analysis) of the cell sizer controlling entrance into S phase, led to the identification of the molecular determinants of the START function and gave insight into their interconnecting regulatory circuits (2). This suggested also that it should now be possible to connect this silicon cell cycle to silicon glucose catabolism, leading to a core blueprint comprehensive of metabolism, growth and cell cycle in a relatively short period of time. For this an integrated activity of the metabolic and the cell cycle groups should be required; the total problem is too difficult for a single groups expertise.

 

SCALE OF AMBITION AND CRITICAL MASS

So far only a few national Systems Biology programs have been launched (USA, Japan, Germany, and Canada) and only one explcit Silicon Cell program (http://www.bio.vu.nl/hwconf/Silicon/). Although the Silicon Cell is a Systems Biology approach (and will as such be part of the large Euroepan Systems Biology Supranetwork ESBIG [http://www.bio.vu.nl/hwconf/FP6/]), it is a highly focused subsection of Systems Biology. The focus stems from the fact that it only calculates on the basis of a complete dataset, without fitting, wihtout theorizing. It calculates what is, and is not directed at finding out what might have been. As such it lacks the focus on the essence of comlex systems, as well as the focus on experimental data other than those that are required to calculate the physiological state.

The making of a replica of an entire cell si impossible for a single scientific grous, as the total expertise required is much too large. We here propose that recognizable parts of the yeast cell will be translated into silicon replica by the very best European group on the topic. Integration of these activities by the IP proposed here will then work towards a Silicon Cell: It is felt that only the focus of the FP6 on strong interdisciplinary research can nucleate the required European network(s). Only in such a way Silicon Biology in this sense of precise replica of living cells, and as a new discipline where Europe is the leader, can be developed. The other Systems Biology programmes in the world lack this focus on making a computer replica of a cell. Those other initiatives do engage in data fitting and approximate descriptions.

The EU already has a long record in supporting research on this organism, since the Yeast Sequencing and EUROFAN Projects.

The aim of this project is to gather the best European groups working on several key aspects of yeast cell physiology in order to develop Silicon replica of those aspects and to then hook these up to each other. It is this strategy that can ultimately lead to the first Silicon cell.

A more immediate milestone of the project is to implement a computer model of a living yeast cell, able to respond to external stimuli by displaying on of the main properties of a living cell, i.e. the coordination between metabolism, leading to increase in cell mass, and cell cycle, i.e. those events required to faithfully replicate and partition the genetic material as well as cell mass. By its very nature the proposed Silicon cell approach allows to tackle detailed description of each module separately, while simultaneously starting the wiring of the system using low resolution versions of each subsystem.

Disassembling of major yeast functions

The separate aspects for which Silicon replica will be developed are given in the above figure.

The making of silicon cells, requires a number scientific activities:

Data mining For each major yeast function identified above, collect available data and try to identify the major regulatory networks, taking special care to the control circuits linking different supermodules.

Iterative development of the model of each (super)module

INTEGRATION

The following organization is proposed for a successful management of the Integrated Project.