Status of the GeneBander tool
GeneBander Working Group
Last update: 10 Aug 00 webmasterGeneBander is an open source web tool freely accessible to everyone. However, if you entered data into the front end, and clicked the submit button, an error message was returned. Why was accessed denied? It was not.
The fact is, GeneBander is only partially implemented at this time. The front end is designed, scripts (algorithms) that do the work are described (1, 2), and the back end product illustrated by many examples.
However the scripts are not suited yet for automatic web implementation. Basically, a friendly and stable shareware text editor (TextSpresso 1.7.1) specialized to the ins and outs of ASCII text strings is used to clean submitted text. TextSpresso is a collection of user-defined filters; a large set of molecular biology filters has been built to clean output from GenBank entries, Blast servers, and the like. The output is then handed to Adobe Photoshop, which then makes the GeneBander graphic gif file that is returned as output to the web server.
Photoshop is needed by the end user in any event to fine-tune and assemble data; it is available for all platforms and comes free with a scanner. However, it is not really required for GeneBander internal workings: the key step is to render strings such as |||.|...|||..|| not as text but as a gif, typically one base pair or amino acid per pixel.
Also, a dedicated stable server is still needed. This will handle incoming requests and process cgi scripts.
How does this all come into reality? That's where the GeneBander Working Group comes in!
Last update: 10 Aug 00 webmasterWho can join? Anybody competent who can help make it happen. There is no registration process. Simply send in your comments and suggestions to the webmaster.
Will I get lots of GeneBander listserve email? : None. There is no listserve. People actively working on GeneBander correspond among themselves. Everything worthwile is collected right here on this moderated web page. All GeneBander news items are posted below in reverse chronological order.
Who has expressed interest? : The following people have expressed interest in possibly helping with various aspects of implementing and improving GeneBander.
Win Hide South Africa Interests: non-commercial, open source, partnerning can only be open collab. James Vincent NIH Interests: programming, scripts, parsing the output of many programs and display it graphically. Gudmundur Oli Hreggvidsson Iceland small biotech Interests: annotating microbial genomes. Juan F. Medrano UC Davis Interests: using it
15 Nov 00 webmasterHere is an example of how to manage GeneBander tracks for web display, with each track being a separate gif stored and sorted according to project in a simple flatfile database or spreadsheet. Of course, by taking a screen shot, the tracks could quickly be consolidated into a single file, say for emailing.
The tracks show basic properties of the prion gene, known and predicted CpG mutations, and domain structure and certain amino acid distributions in the encoded protein -- a comparable display could accompany every GenBank entry. The scale is 1 pixel per 1 nucleotide or 3 pixel width per amino acid. Browsers have no sideways limits on display size, but longer genes would require scrolling at this scale. It is easy to make tracks display vertically by swapping height and width specifications.
These tracks are very conveniently generated from the CDS coding nucleotide subentry at GenBank, as underneath-translated at SwissProt, followed by simple search and replace (note amino acids are distinctively blank-aa-blank). It is convenient to store these search and replace macros in a friendly and stable text filtering application such as TextSpresso.
atggcgaaccttggctgctggatgctggttctctttgtgtactactggagtgacctgggc starting display M A N L G C W M L V L F V Y Y W S D L G atggcgaaccttggctgctggatgctggttctctttgtgtactactggagtgacctgggc ..................|||............|||...|||||||||............ aromatic amino acids now ready for GeneBander track-building
One advantage is that key tracks may need to be displayed several times within a stack to facilitate visual alignments; various subsets of the same tracks might be used in different places according to project emphasis; new tracks might need to be inserted later; a single track might need editing. It is more efficient to individually store all tracks relevent to a given gene and protein in a central folder. Revising a track here automatically corrects all calls to the folder from within a large website.
G+C content G+C blur r=3 CpG all CpG observed CpG CJD sites domain structure non-CpG observed non-CpG CJD all CJD sites domain structure tyr-trp-phe pro arg lys/glu asp
Since a web browser makes automatic accurate vertical height adjustments, these too can be adjusted for the project within a database field, perhaps even by a simple spreadsheet calculation (eg, make total height of graphics stack 100 pixels). Horizontal adjustments are a compromise, the browser shrinks or stretches the display as best it can.
Thus to make the band a little wider (20 pixels instead of 12) and to display vertically at the same scale, simply arrange for the spread sheet to make the necessary changes -- the browser does the rest:
Here are some potential column headers (fields) of the database. After arranging the track database, simply copy and paste desired selection to the html page, removing tabs that separated fields.
-- desired sort order, top to bottom, of graphics -- include in display: yes (1) or no (0) -- general html and track directory IMG SRC="bands/ -- file name of individual track: prnp_GC.gif" -- width in pixels to display: WIDTH="762" -- height in pixels to display: HEIGHT="20" (swap these numbers to display vertically) -- displayed description of track: G+C content -- html code to insure a line break between tracks: br -- non-displayed longer comment or description of track content and purpose