The BIOPEP database developed at University of Warmia and Mazury in Poland is unique in that it focuses primarily on peptides of food origin [17]. It offers the user the ability to generate profiles of potential biological activity of the protein of interest as

well as the frequency of occurrence of bioactive fragments in the protein. For example, in silico analysis was applied to assess the potential of different food commodities to serve as sources of peptides with inhibitory activity against the enzyme DPP-IV, which acts on incretin hormones that play a role in blood glucose regulation Gefitinib in vitro [19]. One limitation is that the DPP-IV inhibitors reported in the literature at the time

of that study consisted primarily of di-and tri-peptides, in contrast to the much longer physiological substrates of the DPP-IV enzyme, GLP-1 and GIP. Higher frequency of occurrence of bioactive sequences in a protein molecule does not necessarily correlate with the potential of that protein to serve as a good source of bioactive peptides unless the potency of each bioactive fragment and any overlaps of bioactive this website sequences are taken into account. To address these limitations, Nongonierma and FitzGerald [20] developed an in silico approach incorporating protein coverage and potency indices, and applied a peptide alignment strategy to investigate the relationship between sequence and activity. Potency is represented in the BIOPEP database by EC50 values, that is, the concentration of the bioactive fragment corresponding to its half-maximal activity. Unfortunately, EC50 values are not always reported in the literature and moreover, may vary for identical sequences if assayed under different conditions. For example

the concentration of a peptide required to inhibit an enzyme to its half-maximal activity (referred to as the IC50 value), can be influenced by assay conditions including enzyme and substrate concentrations. Thus unless the inhibitory activity is reported as the inhibitor affinity constant (Ki), potency of different peptides reported by different researchers may not always be comparable. Molecular docking simulations however have also been applied to elucidate which peptide sequences, either experimentally identified or predicted from bioinformatics investigation, may actually be able to interact with the proteins that are the target of the biological activity [21]. Acharya et al. [22] noted that the dynamic conformational changes induced in both the bioactive peptide and the receptor target protein upon binding impose limitations on computational docking studies, and advocated for a 4D structural database documenting these changes. Nongonierma et al.