From sequence to structure using PF2: improving methods for protein folding prediction

Projects dependent on proteomic data are challenged not by the lack of methods to analyze this information, but by the lack of means to capture and manage the data. A few primary players in the bioinformatics realm are promoting the use of selected standardized technologies to access biological data. Many organizations exposing bioinformatics tools, however, do not have the resources required for utilizing these technologies. In order to provide interfaces for non-standardized bioinformatics tools, open-source projects have led to the development of hundreds of software libraries. These tools lack architectural unity, making it difficult to script bioinformatics research projects, such as protein structure prediction algorithms, which involve the use of multiple tools in varying order and number. As a solution, we have focused on building a software model, named the Protein Folding Prediction Framework (PF2), which provides a unifying method for the addition and usage of connection modules to bioinformatics databases exposed via Web-based tools, software suites, or e-mail services. The framework provides mechanisms that allow users to create and add new connections without supplementary code as well as to introduce entirely new logical scenarios. In addition, PF2 offers a convenient interface, a multi-threaded execution-engine, and a built-in visualization suite to provide the bioinformatics community with an end-to-end solution for performing complex genomic and proteomic inquiries.