The MPR web site http://mpr.nci.nih.gov has undergone multiple improvements his year. There has been another major increase in content, including now 1564 phospho-antibodies specific for 1002 phosphorylation sites on proteins encoded by 440 genes. This has been accompanied by expansion to include antibodies from all the major suppliers of phospho-antibodies and a number of smaller suppliers so that the total of companies is now 12. This has been made possible in part by establishing good communication with major suppliers of antibodies whose antibody annotation was not sufficiently clear to establish unambiguously the corresponding sites of phosphorylation. We have established a dynamic exchange of information from most such companies to facilitate clarification of the exact site, thereby removing ambiguities which impede proper use of those antibodies. This clarification can sometimes markedly expand the usefulness of such antibodies, for example we have made clear to companies (and users) that specific antibodies may also detect paralogs. We are still working with one of the major supplies to clarify specificity of their antibodies. Usage of the site continues to be very strong, with: queries from hundreds of unique IP addresses (from users not search engines) each week. Our tools for analysis of click-through (including retrieval of information on antibodies from web sites to which we link them) demonstrate that most of those users are finding helpful information on MPR. Continued effort on software development is occurring this year and is planned for the coming year. We are well along in a major change in programming infrastructure, namely converting much of the web site to Ruby on Rails (ROR). ROR is a very powerful, open-source framework for developing database-backed web applications. The emerging evidence indicates to us that ROR is on a trajectory similar to that of Linux in its early days, and is becoming the framework of choice for database driven web applications like MPR. We choose to be an early adopter both because of the flexibility ROR it provides us in MPR and because we believe that NCI gaining in-house experience with ROR will be important for facilitating a broad range of application development. We have therefore been worked with NCI Web in setting up an ROR server. The transition of MPR to ROR is mostly completed on a development server and transition on the production server is expected in a month or two. This change will facilitate various aspects of further development harvesting information from the web on phosphorylation sites (which will facilitate the manual curation) and developing a richer set of tools such as end-user comment/curation. Moreover, it allows MPR to use (and contribute to) BioRuby, an open-source integrated environment for Bioinformatics with Ruby language.
