Systems biology aims to quantitatively understand a biological system (i.e., a cell) as a whole. The first cornerstone to achieve a system-wide characterization of the cell was the complete sequencing of the human genome, which encodes the basic information for building the whole protein repertoire of a cell. Next, mass spectrometry-based strategies opened the opportunity to determine the transcribed and translated gene repertoire of the cell - the proteotype. The proteotype bridges the gap inbetween genotype and phenotype and is defined as the actual state of the proteome of a cell. However, to decipher the complexity of biological systems, the definition of the individual components is not sufficient to understand cellular function. One needs to appreciate that a cell is more than the sum of its parts and that biological function is encoded in interaction networks. These protein interaction networks ultimately define protein function and a molecular understanding of such interactions thus enables the analysis of context-dependent cellular signaling. This thesis focuses on proteins exposed to the extracellular space, termed surfaceome, which guide communication of a cell with its outside world. Hence, the surfaceome has a crucial function as gatekeeper, enabling but also limiting cellular communication. Extracellular signals are turned into intracellular signaling responses through the surfaceome in form of ligand receptor interactions. To gain a system-wide understanding of the surfaceome, the identity, quantity and interactions thereof need to be defined. However, most of the available surfaceome information was solely built on the detection of cell surface proteins by a limited pool of antibodies since cell surface proteins were inherently difficult to analyze using other technologies. Only limited information was available about the surfaceome protein repertoire of a cell; a systematic assessement of the variability of the surfaceome over different cell types was absent; only semi-quantitative information about cell surface proteins has been obtained and there was no conception and very sparse molecular knowledge about the interconnectivity between surfaceome proteins. Hence, the aim of this thesis was to identify and quantify cellular surfaceomes, to determine the surfaceome members and to develop technologies for enabling the investigation of the interconnectivity of surfaceome residing proteins. First, the possibilities and the motivation to uncover the biomedical potential of the surfaceome interaction network are discussed in detail within the introductory chapter one, which was written in the form of a review article (chapter 1). i To follow the systematical approach to functionally define a system by first identifying and quantifying its components and then determining its interactions, we set out to first define the identity of the surfaceome (chapter 2). This was necessary, because the available surfaceome maps were limited to the ~300 Cluster of Differentiation (CD) antibody panel. Surfacome sets of 41 different human and 31 different mouse cell types, which were previously collected in a collaborative effort by applying the Cell Surface Capture (CSC, Wollscheid et al, 2009), were used to build the Cell Surface Protein Atlas (CSPA, wlab.ethz.ch/cspa). The combination of these surfaceome datasets revealed nearly 1500 human and 1300 mouse cell surface proteins, which is a five-fold gain compared to the CD antibody panel. Integrated analysis of the CSPA showed that the concerted biological function of individual cell types is mainly guided by quantitative rather than qualitative surfaceome differences. The CSPA is a unique and highly appreciated experimental surfaceome resource demonstrated by 800 monthly website views and an increasing number of citations of the resulted publication. Moreover, the CSPA provided a first blueprint of the interaction space of the surfaceome. To further extend and refine our surfaceome definition, a bioinformatic strategy was developed to create an in silico definition of the surfaceome (chapter 3). Availabe bioinformatics predictions all relied on the same gene annotation databases, which were themselves relying on the limited experimental basis for surfaceome identities, as outlined above. With the CSPA, we had an excellent experimentally validated surfaceome at hand to use as positive training set for a machine learning approach in order to learn characteristic properties of extracellular domains from surfaceome proteins. A model, which incorporated five discriminant biochemical features of extracellular domains of surfaceome proteins was described and then used to predict 2886 potential human surfaceome proteins. On a large cell line panel of 610 cancer cell lines, over 2300 surfaceome genes were found to be expressed in total, with an average of 800 surfaceome genes per cell lines. Interestingly, primary stem cells and their derivatives expressed in average more than twice as many surfaceome proteins. This in silico surfaceome is the first comprehensive and most accurate definition of the surfaceome and is the basis for all future surfaceome interrogations. This resource is available under wlab.ethz.ch/surfaceome, which also provides user-based visualisations of surfaceomes. With the CSPA and the in silico surfaceome, the global and cell type specific identity of the surfaceome was defined, accomplishing the first step towards a system-wide understanding of the surfaceome. ii The next step in the systemic assessment of the surfaceome was to investigate the surfaceome interaction network. Since current protein-protein interaction (PPI) technologies were hardly applicable to cell surface proteins, it was necessary to develop and tailor PPI technologies to specifically target the surfaceome. The concept of radical based biotinylation was applied and combined with hydrazide chemistry to first functionalize and than target glycans at the cell surface. The Proximity Radical Tagging (PRT) technology was established in order to reveal lateral surfaceome interactions of specific receptors (chapter 4). Proof-of-concept applications, like the detection of proteins associated with lipid rafts and heterodimer partners from Erbb2 and the toll-like receptors demonstrated that PRT is very sensitive and able to reveal proximity information of surfaceome members (chapter 4). PRT was then used to investigate the nanoscale organization of the surfaceome in a larger scale on different cell lines (chapter 5). Several cell surface proteins were targeted by PRT and evidence was found that certain lateral surfaceome neighbourhoods change between cell types, whereas other interactions are stable. It was further demonstrated that the strength of the interaction and probably also a distance constraint could be revealed by appropriate experimental setup. The PRT technology opens for the first time the possibility to accomplish the last step of a systemic elucidation of the surfaceome. Further interrogations of the surfaceome interaction network will reveal new mechanistic and functional insight into the surfaceome with crucial implications for the development of novel therapeutics, as for example multitarget drugs. In summary, the first experimental and bioinformatic definition of the surfaceome was achieved, outlining the protein repertoire of the surfaceome. Further, a tailored technology for the investigation of the lateral nanoscale organization and the interconnectivity of the surfaceome was developed. The CSPA and the in silico surfaceome are both provided as tools for the community, for the rediscovery of surfaceome proteins. The PRT technology allows for systematic, large-scale surfaceome interaction screens in order to further elucidate the functional consequences of dynamic nanoscale changes at the cell surface.