Known Bugs

Logs & Bugs

Goto: Administrative Log | Bugs

Administrative Log

December 15, 2007
1) The address of the PFAM ftp site has been changed to "http://pfam.sanger.ac.uk/help?tab=helpFtpBlock". The DRAGON program has been updated to reflect this change.

July 11, 2005
1) We have redesigned the front page of DRAGON.
2) We continue to update the DRAGON sources once each week. DRAGON incorporates 52 UniGene species.

March 24, 2003
1) We have integrated the following 13 new species from UniGene into DRAGON: Caenorhabditis elegans (Cel), Ciona intestinalis (Cin), Chlamydomonas reinhardtii (Cre), Dictyostelium discoideum (Ddi), Gallus gallus (Gga), Glycine max (Gma), Lycopersicon esculentum (Les), Medicago truncatula (Mtr), Oryzias latipes (Ola), Sorghum bicolor (Sbi), Sus scrofa (Ssc), Silurana tropicalis (Str), and Solanum tuberosum (Stu). All files in DRAGON continue to be updated every Sunday morning at 3:30 AM.

August 23, 2002
1) We were having some issues with the index buffer for MySQL when trying to load the UniGene database into DRAGON. This issue persisted despite increasing the memory allocated to indexing. As it turns out, this is a bug with the MySQL version we were using, 3.23.21. Upgrading to 3.23.52 seems to resolve this problem.
2) The UniGene files in DRAGON now contain all organisms available in UniGene. These include the animals: Anopheles gambiae (Aga), Bos taurus (Bt), Danio rerio (Dr), Drosophila melanogaster (Dm), Homo sapiens (Hs), Mus musculus (Mm), Rattus norvegicus (Rn), Xenopus laevis (Xl), and the plants: Arabidopsis thaliana (At), Hordeum vulgare (Hv), Oryza sativa (Os), Triticum aestivum (Ta), and Zea mays (Zm).
3) The DRAGON database is upto date as of August 22, 2002. DRAGON will be automatically updated every Sunday morning at 3:30 AM to 5:30 AM Eastern Time. Do not plan on using Annotate or Search during this time, as your querying will be extremely slow and innacurate.

August 9, 2002
1) Choosing to return Kegg pathway numbers from Annotate is much faster now. Also, if you choose HTML output, the hyperlink on the Kegg numbers will take you to biochemical pathway diagrams, displaying where in the pathway your gene is situated. This is similar to what Dragon Paths does, with the exception that the link in Annotate will only display one data point (one gene) while Dragon Paths will show all data points with the same Kegg number. The Kegg datafiles are upto date as of August 8, 2002.

July 31, 2002
1) Fixed some performance bugs with the Dragon Search tool. Also, included more hyperlinks in the HTML output.

July 29, 2002
1) HTML output on Annotate now includes a hyperlink for the cytoband return field. This link will take you to the Human or Mouse Map Viewer at NCBI and give you a chromosomal display of the gene in relation to neighboring genes. Note, that there will only be hyperlinks for human and mouse genes.

July 26, 2002
1) Drastic improvements made in speed of Annotate. Working on making querying the Kegg database faster.

July 24, 2002
Major Upgrades to DRAGON
DRAGON is being maintained and upgraded by Vinoy Prasad. 1) Fixed a major bug with DRAGON Annotate. For a given GenBank Accession number, if either Unigene return fields were missing, or SwissProt return fields, or both, the entire row would be deleted. Annotate now does what it was intended to do; that is, for an input of GenBank numbers, all data points are outputted with their corresponding return values or lack thereof. DRAGON will no longer delete rows in your data set.
2) Improvements to the Annotate html output. Most of the return values have been hyperlinked to their corresponding Unigene, Swissprot, PFam, Entrez, and PubMed sites. Additionally, we added borders to the HTML table and made the output of amino acid sequence more organized.
3) The Medline return field has now been changed to PubMed.
4) For GenBank numbers with no Unigene entries or GenBank numbers that have been filtered out, Annotate will allow you to view SwissProt and PFam returns if they exist. Certain GenBanks are large fragments that span multiple genes, and, thus, can have multiple SwissProt entries associated with them. Annotate will output all such entries, but there are different ways to view them. By choosing to output "All values on one line", the SwissProt and PFam return values will be prefixed with a 1> 2> 3> etc. corresponding to the different SwissProt IDs. Choosing to output as "Multiple rows, one value per row" will not give you these prefixes. Outputting multiple rows is best used if the output will be used in the DragonView tools, but is not recommended if you will be simultaneously returning fields that can have multiple return values, ie Keywords and PubMed numbers together, as the number of rows for each data point will balloon.
5) If your microarray data has version numbers appended to the Accession number, eg NM_000970.1, you can still annotate your data with DRAGON as version numbers are ignored.
6) All databases are upto date as of July 21, 2002.
7) We will be adding the Ensembl database to DRAGON in the coming weeks.

March 15, 2001
1) Provided borders for the Search html output and other minor site work.

February 21, 2001
1) Changed the field delimiter on the Annotate page back to tabs.

February 16, 2001
1) Made the default field delimiter on the Annotate page commas instead of tabs.

January 18, 2001
1) Made design changes to site.

December 11, 2000
1) Optimized MySQL memory buffer allocations in order to speed-up the annotation of large uploaded datasets.
2) Made it possible to use the same filename for uploaded files over and over again.
3) Fixed a glitch in the uploads section that was preventing annotation of all possible Genbanks.

December 4, 2000
1) Made file uploading feature available on the Annotate page.

November 27, 2000
1) Implemented redundancy filters for the html and text-based outputs from the Annotate page. You won't get any identical lines of output from now on.

November 21, 2000
1) Provided for hyperlinks in the html output of the Annotate and Search pages. Found a bug in the Annotate page where the annotated columns are shifted down one row if both Swissprot and Pfam are used. Working on it.

November 14-15, 2000
1) Worked out some bugs in Dragonview and spruced up the web pages some.

October 20, 2000
1) I fixed the redundancy filter for the human unigene 5 file. This prevents the inappropriate association of certain Genbank numbers with other sorts of information. Specifically, the filter deletes any Genbank numbers that are associated with more then one Unigene cluster. These Genbanks are usually associated with large fragments from cosmids, YAC's, BAC's etc. Since any Genbank that is pasted into the Annotate window is linked up with all of the other Genbanks in its associated Unigene cluster, if that Unigene cluster also contains one of these large Genbanks, the Genbank that was pasted in then gets associated with more then one Unigene cluster and as a result with any other information associated with both of the Unigene clusters. The redundancy filter will soon be run on the other species in the Unigene as well.
2) I added some more links to the links page and updated the index page a bit.

August 28, 2000
1) The KEGG numbers should be working now for human genes but not other species. Some other bugs have been noted recently. Here they are:
- The text box on the Annotate page will only accept a certain amount of text! We're working on an upload feature.
- If a gene does not have a KEGG number associated with it, you won't get any info back on that gene, even if you have asked for other sorts of info on it.
- The redundancy filter seems to be on the fritz. This means that you could get the incorrect association of Unigene numbers with Swissprot numbers for some Genbank numbers. We're working on it.

July 24, 2000
1) I think that we now have the annotate page running fast again too. So, now the search page and the annotate page should spit something back pretty quickly for reasonably sized data sets (< 1000 Genbank numbers). Remember that for large data sets (> 5000 genes) try annotation first with html output. However, if things go slowly then try for an email output and check to see how long the search took at the bottom of the file when you receive the email.
2) The names were taken out of the example data set. White space characters and other non-alphanumeric characters were causing a strange error for some people. The example data set is simply a list of Genbank accession numbers now.

July 20, 2000
Well, we got the search page fixed, so that it is running much faster, but now the annotate page is running quite slow. For example, it took 183 seconds to run one Genbank through all attributes from all databases. It used to be that you could run 40 Genbank numbers against all attributes in under one second. I think that we can fix that tommorrow. We also are now running on the new server! Woo-Hoo! It's a shiny new Dell Poweredge 6300 with dual Xeon 550 MHz processors and 2 GB of RAM running Redhat Linux 6.2. In the process of migrating to the new server other things may have been broken. So, we'll hope for the best and see what happens. Cheers!

If you have questions, comments or have noticed other bugs...
Please email either Jonathan Pevsner (pevsner@kennedykrieger.org) or Christopher Bouton (bouton@bu.edu). We appreciate your feedback.