NMPDR Development Blog

I have finished the design of the drug target section of the Sprout database. (The two sections connect via the Feature entity.) You can view the design diagram here. Clicking on an entity or relationship will bring you to the corresponding section of the database documentation.

The drug target data has been loaded into Sprout. You can review the database design here. I coded a quick display script so you can see what the data looks like. (PLEASE NOTE: It's just a data dump so that people can critique the data design, not a finished display.)

Yesterday Matt, Jen, and I put up some new drug target pages on the Development Web Site. The entry point for the new pages is http://web-1.nmpdr.org/next/content/docking_proof_of_principle.php. The entry page shows the three PDBs used as examples in Gohar's thesis, and contains links to some charts and to a page that displays the ligands and the docking results.

The ligands page is generated from an updated version of the ligands.cgi script from the bio servers. The major change from the original script is that we show 50 ligands at a time instead of all 10,000 on a single page. (The latter ran into problems due to the template philosophy used elsewhere in Sprout. So that the best results were shown first, we used a utility to sort the data files so the examples with the lowest binding energy appear first.

Hopefully, this will be the first of many pages of drug target results we can show on the NMPDR.

A drug target search is now available on the NMPDR. This new search allows you to select one of the drug trarget projects from a dropdown menu and display all the features found by that project. The search styles have been changed to make the results tables use the alternating-color scheme used elsewhere on the site.

The drug target search page can be found at http://web-1.nmpdr.org/next/FIG/SearchSkeleton.cgi?Class=DrugSearch.

Please let me know if you feel we need more or different columns, or even additional searches for drug target data.

The NMPDR will be using a new approach for the drug target project beginning in version 17.

Instead of working tables, the drug target data will be divided into three categories: Drug, Vaccine, and Toxin. The data will be curated manually before it is presented to the NMPDR to insure that the PDB associated with a PEG is a qualified match.

Version 17 will start loading Friday evening the 17th. At that time, three flat files (one per category) need to be available at /disks/nmpdr/v16/FIGdisk/FIG/var/DrugTargets so they can participate in the load.

A new script has been written on the bioseed server that allows users to focus on data for a particular organism or a particular category. This script will be adapted for NMPDR and then linked into the drug target pages. The new script will go live on December 1.

Version 17 of the NMPDR is now available on the development server. It includes the new drug targets pages, which can be seen at http://web-1.nmpdr.org/next/FIG/targets.cgi, though the information there has not yet been completely curated.

Not all of the attributes we want are available. Once the attribute system is fixed (hopefully in a day or two), I will reload the NMPDR property table.

I have created a new script, DrugClean, which prepares a drug target output file for NMPDR. The script is invoked as follows:

The macFile switch is only necessary if the input files are all in Macintosh format.

The script will remove duplicate entries and entries for PEGs that are not in the current version of the Sprout database. It also converts the file to Unix format. I have changed targets.cgi to expect Unix files, so when we get new drug targets files this script must be run on them or they won't work.

I also fixed a performance problem with the organism files, and they now load in around 10 seconds instead of 50.

Some improvements have been made to the drug target pages in version 17.

• The performance of the organism lists has been greatly improved. Previously, the category for a peg (drug, toxin, or vaccine), was determined by searching the category files for each peg. Now the category files are read once and the category information is kept in memory.
• There is now a column for a peg's best hit to the human genome. Only hits with a score of 1e-15 or lower are shown. The column contains the hit score, and clicking on it will bring up the protein page for the specified human peg.
• The header columns no longer appear when there is no data to display.
• The template has been changed to a wide display so more data can fit on a screen.

Because these are version 17 changes (version 18 doesn't exist yet), they were slipstreamed into the live NMPDR.