Proteins are nanosize objects made of amino acids in interactions. As all naturel systems, proteins are robust, adaptable and only rarely fragile. In order to understand the mechanisms of robustness to mutation we investigate the topology of adjacent amino-acid networks associated with proteins. We first compute assortativity measures on these biological networks and compare the results to the assortativity of various networks (power grid network, financial networks, gene regulation network, …). In a second time we study the robustness of these biological networks and compute some similarity measures. We discuss the interpretation of similar amino-acids in terms of link redundancy and we show how error correcting codes are implemented directly by the evolution in proteins.