Comparing two protein structures by FATCAT. Users can provide two protein structures by uploading a pair of files in PDB format, or inputing a pair of PDB codes. FATCAT then calculates their structure alignment using either rigid or flexible comparison, depending on the users's choice.

Searching for structures similar to a query structure. One can either upload a structure file in PDB format or input a PDB code as the query. The databases the query structure will search against are SCOP and PDB at different redundancy levels(Sel40 and Sel90).

NOTE: because of the massive calculations in comparing 3D structures, it usually takes a few hours for one query before one gets the results.

TIPs: If your query is from PDB or SCOP, you can preview the results for your query in our precalculated database( Structure Neighbors). It usually takes only a few seconds or less. Most of the times, you may not need to run the Database Searching again if your query is precalculated.

A database of precalculated structure alignments for proteins in both PDB and SCOP, using flexible FATCAT. The database is searchable by any given PDB or SCOP code. The main reason of this database is because the live "Database Searching" takes a very long time, although the "structure neighbor" server provides many other convenient features.

For details about What and How of the "Structure Neighbors" server, please read FATCAT Structure Neighbors - help page.

1. What is the difference between "Database Searching" and "Structure Neighbors"?

   In short words, both "Database Searching" and "Structure Neighbors" are for finding structures(from a structure database) that are similar to a query structure. But they use different implementation strategies and both have something good or bad comparing to each other.

   "Database Searching" is more flexible but more time-consuming comparing to the "Structure Neighbors" server. The user can input a arbitrary query structure as long as it is compatible with the PDB file format, or simply input a PDB code. The structure databases in which the query will find its similar structures are larger. Right now, we provide both PDB and SCOP in SELECT-40 and SELECT-90, and it can easily add more structures to the database if necessary. However, because it calculates the structure alignments between the query structure and all the structures in the database after the user has input the query, it usually takes a very long time(~ 10 hours) to have the results back to the user. Except for its long waiting time, "Database Searching" is for more general cases.

   "Structure Neighbors" is a database of precalculated structure alignments. Because it's precalculated, it's very fast. Uaually it only takes a few seconds or less for a query to find its similar structures. For the same reason, it is, however, not as flexible as the "Database Searching". The user has to input either a PDB code or a SCOP code as the query structure, rather than a arbitrary structure, and the structure database that a query can search against is fixed (please see here for details). The structure database includes both SCOP(PDB structures in domain format) and NEW_PDB(PDB structures in chain format; NEW_PDB means PDBs that are new, i.e. not annotated by SCOP yet). The all-by-all structure comparisons have been calculated on SCOP_90, NEW_PDB_90 and both. Although it is not flexible, it included the most frequently used protein structures. It also provides a lot of other conveniet functions, such as graphically mapping classified domains into a PDB protein. Moreover, because the "Structure Neighbors" is built on the top of a relational database, it is much easier to change the result set. Most importantly, it included the structure neighbors for all proteins in PDB up to the update date of our server. We try our best effort to keep up with the PDB site.