In the last 15 years, the study of protein aggregation has evolved from a mostly neglected topic of protein chemistry to a highly dynamic research area which has expanded its implications through different fields including biochemistry, biotechnology, nanotechnology and biomedicine. The analysis of protein aggregation has attracted a particular interest in the biomedical and biotechnological areas. Because, on one side, the formation of insoluble protein deposits is associated to an increasing number of human disorders, many of which present fatal pathological consequences. And on the other hand, aggregation is a frequent shortcoming in the recombinant expression of proteins at the industrial level, such as in the production of proteinaceous therapeutic agents like antibodies. Consequently, the survey of mechanisms to prevent protein aggregation is currently the focus of deep investigation with the aim to develop preventive or therapeutic methods for the intervention of these depositional disorders and to enhance the yield in the biotechnological production of proteins. The power of the computational tools developed to predict protein aggregation has fostered the identification of the determinants influencing the aggregation of polypeptides and has allowed to investigate how the selective pressure to avoid aggregation has shaped the cellular proteomes along evolution. From these analyses, different mechanisms to prevent protein aggregation have emerged ranging from negative design strategies found in polypeptide sequences and structures, to the characterization of the factors governing the cellular machinery in charge of the protein quality control. The present thesis provides a multiperspective analysis for the detailed characterization of several of these mechanisms evolved to confront the risk of protein aggregation. In this sense, the use of a variety of specific proteic models of aggregation or different sets of proteins with related properties has allowed to analyze particular strategies to avoid aggregation in depth. At the same time, the approach based on the study of closely related ensembles of proteins has allowed to identify functional constraints that limit the evolutionary selection against aggregation. More specifically, the work presented here addresses the effect of restricting the configurational freedom of the polypeptide chain by disulfide cross-linking on the aggregation process, as well as the impact over this phenomenon exerted by the presence of intrinsically disordered protein regions. Additionally, the regulation of cellular protein abundance as a function of protein aggregation propensity has also been surveyed. On the other hand the analysis centered on the study of closely related proteins has revealed how the requirements to fold efficiently and to maintain the catalytic activity constrain the minimization of the aggregation propensity of proteins. These analyses highlight particularly the interplay between folding and aggregation, in such a way that the analysis of the aggregational properties of polypeptides allows to forecast the mechanism of folding of certain kind of proteins.