In 2000, our Fold and Function Assignment System (FFAS) server pioneered protein profile-profile algorithms, applying them to protein structure prediction. Since then, the basic ideas underpinning these algorithms have been used in over 20 distant homology recognition algorithms, and the public FFAS server is used by almost 1,000 registered users running hundreds of jobs per day, applying the results not only in protein structure prediction, but also in function prediction, target selection in structural genomics, and general analysis of diverse protein families. With ever increasing flow of new protein sequence, many of them representing new, uncharacterized families, the importance of distant homology recognition is constantly growing. We propose enhancing the FFAS structure and interface to match these new types of applications, developing the server into a major resource for studying broad and diversified protein families. We plan to extend the usefulness and maintainability of FFAS by restructuring its code using modem programming practices to develop a modular, multistage program ready to be integrated with other servers, as well as use other programs, developed both inside and outside our group, to improve quality of data at each step of the prediction process and also to export intermediate results to the user for analysis. By extending the set of analysis and visualization tools integrated into the FFAS server and improving its user interface, we want to make it easier to be used by a generally trained biologist. Finally, we plan to perform a significant hardware update to avoid delays in providing annotations for user-submitted sequences.
Provide annotations for uncharacterized proteins, including human disease-related proteins and virulence factors, to help understand their functions and thus shorten discovery timelines for new therapies.
Showing the most recent 10 out of 12 publications
