MDB News

This is a minor release. The next major release (version 2.0) will include more functionality and a better data set (automatically recognized first and second shell, nuclearity, geometry, etc.). Until the next major release updates to the data content will be made, but the recognition algorithm used in the versions 1.x will not be changed.

• The database contains the metal-binding sites from entries in the PDB database as of December 03, 1999: the MDB about 3700 macromolecules, with a total of about 10700 metal-containing sites.
• A script demonstrating our idea for a Web-based MDB API (more information on this topic will come later). The script accepts a valid "SELECT" SQL statement, and generates a result set in one of three different formats: HTML table, Comma Separated Values or as a WDDX packet. Go to the MDB API page for more information. [beta]
• A series of other search and display interfaces have been added in a simple implementation. Full implementation of these capabilities will be done once the next major release (version 2.0) is completed. Go to the MDB beta test page to use the:
  1. Metal-ligand histogram/statistics calculation form
  2. Metal liganding pattern plot form
  3. Examples of XML/CML representations of metal-sites
  4. Chime examples of metal sites containing first and second shell information.
• Fixed the problem with extraneous ligands showing in the display and in the "mini-pdb" files, due to non-unique residue numbers in some PDB files.
• The "ligand files" format has changed, the information in the "Packages" page will be updated to reflect that.
• The advanced forms have been modified to allow for more complex searching.
• A bug in the selection of year ranges when using the advanced form was fixed
• Problems with the protein descriptions has been solved.
• Search forms now check that the user send some reasonable combination of input. An appropriate error message is now displayed if some critical piece of information is missing or if one of two linked conditions has not been set.
• To avoid infinite time searches, the search scripts have a limit of 3 minutes of running time, this does not include the time taken by the database server itself.
• Also to avoid the problem above, the database server will refuse to make a search in which the intermediate table (product of joining two or more tables) will contain more than 1'000,0000 (one million) rows.

• The site information contains only first shell ligands, and does not automatically recognize multinuclear or cluster sites. The next major release solves this problem (see the beta page for a preview).
• Hemes contain sometimes extra ligating atoms, because the algorithm for the version 1.x has some problems with recognizing automatically those systems.

• A clarification. In the search forms in this site, "number of ligands" refers to the number of atoms ligated to a metal center. The number of residues ligating to a metal will then be equal or less than that number.
• The "Edited" database will no longer be updated. The indexing algorithm used in constructing the next major release (version 2.0), will make the information in the "Edited DB" redundant. This part of the MDB has been frozen since release 1.2 and will remain so.

Changes:

1. Search interface additions/updates:
   • Added new serch interface ("Advanced Queries") that allows more complex search of the MDB. Three forms are offered:
     • One that searches using PDB ids
     • A simplified form similar to the one included in the interactive search interface (i.e. the one containing the viewer applet)
     • A more complex form that allows more interesting searches to be made on the MDB.
     Look for more info in the "Advanced Queries" page.
   • Modified the navigation, layout and options in the interactive search interface for both the "Raw" and the "Edited" databases
2. Back-end (Query processing, Database, Site documents) changes:
   • Minor corrections in the r-values, resolution and description of the metalloprotein entries to reflect the revisions up to January, 1998.
   • Optimization of complex queries parsing to improve response.
   • Allow searching on deposition/revision dates, and other numeric fields by using a complete range syntaxis.
   • Modified the whole site to support SSI and on-the-fly generation of content, to make extensions and modifications simpler.

• It is now possible to indicate the maximum number of hits to return.
• Words in the authors names can be used to restrict the search.
• A series of "canned" extra options that allow a more complex search. Click on the [More...] button in the search form to access the options. These options will let you require that:
  • the site contains at least one water as ligand
  • the site does not contain any nucleic acid as ligand
  • the metal-ligand distance is less or equal to 3 Â
  • the PDB containing the site has been revised in the last 2 years (1997-1998).
• A "remote-viewer" interface has been implemented as a way for people to refer to sites in particular metalloproteins. Select the links above for more information.

The pages in this site are being updated also to reflect the changes in the MDB. This is a stable beta release and we expect that the final release will be up soon.

• The "raw" database is now in SQL format (mSQL as database engine)
• The "raw" DB now contains information from all the released PDBs as of September 15, 1998.
• We are using PHP3 to perform the searches, which combined with the mSQL server gives us better performance, and the possibility of more complex queries.
• IMPORTANT: in the new search interface you have the option of requesting to be shown all the hits found, if you do not check that box the search will default to showing the first 30 results as ordered by their respective pdb code.
• If you do not want to search on Resolution or R-value, input "NULL" in the respective boxes.

We have tested in-house this new release, but bugs have a way of creeping out when you are not looking, so we will appreciate your input if you find one. To help us debug, the queries submitted are being kept as SQL commands.

Also, the FAQ page has been updated to include information on how to reference the MDB, and the URL's to use.