Bioinformatic Research In Proteomics

A bioinformatics platform was been needed to deepen the researches of proteomics and spider toxinome launched jointly by our laboratory for many years, which including the local combined database platform, the perfect integrated system of software, new calculation theory, new algorithm and new software system developed independently to deal with and analyze the initial data and to obtain the results that can't be received by traditional biological experiment. In the present study, three works were conducted: first, set up the merger databases of mouse and human; second, develop a comprehensive analytic platform including information retrieval from the result file, prediction of protein function and protein-protein interaction, metabolic pathway reconstruction; third, refine the phylogenetic profiles method used to predict protein-protein interactions and develop a method used to Metabolic reconstruction based on sequence information, which systematize the process of data processing and data analysis.The availability of complete genome sequences now inspired the development of tools for functional biology on a proteomic scale. Several experimental approaches or in silico algorithms have aimed at clustering proteins into networks with biological significance. Among those, the phylogenetic profiles method, one of popular in silico proteomics methods, has been widely used to predict protein-protein interactions. Recently improved versions were developed and applied at a genomic scale to reconstruct the protein interaction networks and discover new cellular systems by Date et al~1 Here we refined the phylogenetic profilesmethod based on above improved versions, which integrated information of organisms' phylogenetic relationship, homology finding and protein's functional annotation to improve the prediction accuracy. The investigation of reference organisms and homology finding in detail showed that they have critical effect on the prediction accuracy of the phylogenetic profiles method. Comparison our protein interactions data sets with those predicted by above improved versions revealed that the refined phylogenetic profiles method described here is more reliable for the prediction of the protein's functional linkages.We have also made a thorough study on the method of biochemical pathway reconstruction. Elucidating the biochemical pathways and other types of protein interactions can enable systematic interpretation of sequence data and suggest testable functional assignments for hypothetical proteins. Several approaches have been developed to measure protein interactions and to reconstruct biochemical pathways of whole species, but most of them have certain limitation. Transferring conserved metabolism pathway information from well-studied species to a newly sequenced species is of significant value. Therefore, we investigated the extent to which metabolic pathway information in one species can be used to reconstruct pathways in another species. We used interologs, information transferred from the interaction data of a source species, to search for the "metabolog" in the target species. The metabolog is assigned to the same EC number or is predicted to have the same enzyme action, with the relevant protein, in the source species. From the highly accurate interaction data of saccharomyces cerevisiae, which contains 3462 interactions among 1414 proteins and 52 reconstructions of biochemical pathways, we identified 9628 potential...