Insights into the EFSA protein risk assessment: in silico tools : Protein toxicity: an integrated in silico pipeline

Several proteins are associated with adverse (toxic) effects in humans and animals, by a variety of mechanisms. These are produced by plants, animals and bacteria to prevail in hostile environments. In the present report, we present an integrated pipeline to perform a comprehensive literature and database search applied to proteins with toxic effects. “Toxin activity” and “toxin-antitoxin system” strings were used as inputs for this pipeline. UniProtKB was considered as the reference database, and only the UniProtKB curator-reviewed proteins were considered in the pipeline. Experimentally- determined structures and homology-based in silico 3D models were retrieved from protein structures repositories; family-, domain-, motif- and other molecular signature-related information was also obtained from specific databases which are part of the InterPro consortium. Protein aggregation associated with adverse effects was also investigated using different search strategies. This work can serve as the basis for further exploring novel risk assessment strategies for new proteins using in silico predictive methods.