ABSTRACT
Systems biology relies heavily on a number of preliminary steps for preparing high-throughput experiments and making the results readily available for biological analysis and modelling. Though these steps are not per se part of what we commonly define as systems biology, they are essential for enabling the systems biology approach (Ghosh et al., 2011). Therefore, this chapter presents an overview of bioinformatics tools and standards used in a typical analysis workflow Figure 4.1 which includes the following steps. Once the biological and/or clinical question is posed (Ê), an experimental design is defined in order to efficiently answer the problem raised (Ë). Then, the high-throughput experiments are performed (Ì). A scanner generally analyses the microarray∗, sequencing slides or phenotyping screening, and produces images which are processed using appropriate algorithms to quantify the raw signal (Í). This step is followed by normalisation which aims at correcting the systematic sources of variability in order to improve the signal-to-noise ratio (Î). The quality of data is checked at the level of both the image analysis and the normalisation steps (Ï). At this stage, the information provided after normalisation is still rough. The meaningful biological information relevant for biologists must be extracted from the data (Ð). Once the relevant information is extracted, the data can be used in a transversal analysis to perform clinical biostatistics, classification or systems biology approaches (Ñ). Finally, the results need to be validated, interpreted and can lead to new experiments (Ò). The bioinformatics workflow and computational systems biology approach are cyclical processes involving data acquisition and preprocessing, modelling and analysis. The integration and sharing of knowledge help to sustain the capabilities of this cycle to predict and explain the behaviour of biological systems. Therefore, to be successful, the workflow strongly relies on enabling processes to annotate (À), manage (Á) and compute (Â) the data. In this chapter, the steps Ë, Î, Ï and the processes À, Á and  will be described. Steps Ð and Ñ will be raised from Chapter 5 to Chapter 12. The image analysis will not be addressed in the present book but the reader can refer to Fraser et al. (2010) and Novikov and Barillot (2007). Finally, this chapter illustrates how knowledge from the literature and databases can be extracted, and visualised using appropriate standards and software used in computational systems biology.
