1
§ The clickable image of the complex
§ Links
o The protein pages themselves
o Suggestions for automation of this procedure
The clickable image of the complex was created in a UNIX environment using a cgi file, clickimg.pl, created by Jack van Rooij. Click to download. This program creates the image file (in gif format), and also the HTML code (in a html-file) which is needed to make the image clickable. To make these two files the program needs a data file (*.dat) as input. Click to download. This file contains the coordinates and hexadecimal color code of each individual sphere. The first sphere mentioned in the data files is drawn first by the program, so spheres defined afterwards in the data-file will lie over this sphere if they happen to overlap (depending, of course, on the input coordinates). The input coordinates are defined in the following order: ‘x-axis coordinate of the center’, ‘y-axis coordinate of the center’, ‘width’ and ‘height’. For example, the first sphere (SWI2) is defined by the following line;
98, 60, 105, 47, #CC3399, SWI2
The code #CC3399 represents the hexadecimal colour code and SWI2 represents the name in the center of the created sphere which will be clickable. Hexadecimal colour codes can be found on the internet using a search engine (I suggest using www.google.com). The colours were chosen such that the protein names would be properly visible (i.e. light colours)
clickimg.pl clickimg.dat GifFilename.gif > HtmlFilename.html
Where you can specify the desired Gif and Html filenames. After running th program an error message will appear (“Bad free() ignored at clickimg.pl line 98.”) but this is not significant (the -writer of the program does not know exactly why this message appears, but it is clear that it is nothing critical. Besides, the program works, so who cares, right?).
The program has now created the html-code for the clickable map, in the html-file. This can be cut and paste into a html document or the created file itself can be made into a real html-file by adding the appropriate tags.
The following HTML-tag attributes were altered manually after creation of the files:
(Optionally, the place where the new page loads into can also be altered by changing the TARGET attribute, for example to “_new” which will open the page in a new window)
The links in the main page where found with a google search, and the most useful pages were selected.
Protein names and Sequence Length were copied from the Swissprot files. YPD seems to have the most complete list of Gene Names so these were copied from there.
Most of the links in these pages where retrieved directly from the Swissprot files by copying and pasting their html-code. With the aid of Word text processor the yellow colour which the table has in these Swissprot files was changed to white by replacing the appropriate hexadecimal colour codes, i.e. <tr bgcolor=”#FFFFCC> with <tr bgcolor=”#FFFFFF”>. But this is not strictly necessary as Swissprot files are public access.
The link to Swissprot was added manually, and the following links were found in other databases and added at the bottom of each table. (The database name in between brackets is where each link was found).
The following databases contain links which could also be incorporated in future:
All protein sequences can be found here for fast copying of all sequences for a BLAST run which will be needed every six months or so to keep this data up to date. These sequences were copied from the YPD database.
BLAST’s of the proteins against the PDB database were performed using the CMBI BLAST2 server. Standard settings were used (unless stated otherwise) i.e.. Filter ON, Word size 3, BLOSUM 62, GapOpen/ExtensionRepair 10/1
Note that BLAST results can also be found in GeneQuiz.
In my view it is possible to
automate the creation of analogues reference pages. Within the CMBI a script
exists which was used to create the GCPCR Database. This could be modified to
include the links which are not in Swissprot, mentioned earlier (see Database links, Suggestions).
Hopefully in future the Genequiz database will contain stable instead of
unstable links so this can be linked to permanently also.
It should also be possible to automate the BLAST runs to run every few months or even every day, if necessary. I think BLAST server capacity must be kept in mind though.
I suggest making database pages open in new windows or better still in frames. This way it will be easier to keep track of the protein you are working on, as the browser-window will still show the name of the protein in the task bar. Also it means you don’t need to use the ‘back’ button so often!
Created by Frank van Zimmeren, 28/9/01