11/01/99: EpoDB Server is now back to function following an extended close during which our system was reconstructured. We apologize for any inconvenience the close may have caused you. We would also like to alert you that the bioWidgets in EpoDB is not fully functional at this time. If you have noticed that EpoDB has not been updated lately, you are right. We are currently integrating EpoDB into a new database named HemoDB. We anticipate HemoDB to be released shortly and we will make the announcement here and provide the links when it happens. Thank you for using EpoDB and please don't hesitate to send your comments and suggestions to epodb@pcbi.upenn.edu

4/23/99: The "Recent paper on EpoDB" has been replaced with a page of papers on EpoDB. Papers in the 1998 and 1999 Database issues of Nucleic Acids Research can be downloaded as pdf files.
The TESS server that EpoDB uses is now www.cbil.upenn.edu which should speed up response.

2/22/99: Attention MacUsers. A new Mac OS Runtime for Java (MRJ) 2.1 is available from Apple. This may allow the bioWidgets to work better for you.

1/25/99: BLASTView has been updated to search libraries containing the latest EpoDB entries. BLASTView now contains an "Analyze" option which shows protein motif annotation from the BEAUTY server at Baylor.

11/4/98: The new entries in EpoDB v2.2 are now available through most of the queries. Still to come is including the new entries in the Gene Groups query, the BLAST search, and the graphical view. The number of new entries is given on the "keyword" page and the new names are given on the "controlled vocabulary" page.

Three major advances have occurred in the 2.2 version of EpoDB.

• The database has been materialized in Sybase relational tables in addition to Prolog. This will make it easier for us to provide queries that access more of the data in EpoDB and to write complex queries that use more than one data source. A first example is the new Gene Expression query.
• A BLAST-based algorithm has been developed to group GenBank entries representing the same gene. The transcribed sequences of the entries are used and entries are grouped together if the p-value of the match is e-50 or smaller and the sequence identity is at least 98%. This allows more of the entries to be accessed directly using a controlled vocabulary for gene names. The Gene Group query provides this function along with the ability to search using keywords (an uncontrolled vocabulary).
• The sequence similarity protocol used to organize GenBank entries has been applied to updating GenBank entries. A keyword search of GenBank release 107 returned nearly 15,000 new candidate entries for EpoDB. Instead of manually reviewing each for inclusion into EpoDB, the sequence similarity protocol was used to identify those entries with p-value matches to current EpoDB entries of e-50 or smaller. Repeat sequences were masked and only transcriped sequences were used from the current EpoDB entries. The result was 1407 entries consisting mainly of orthologs to EpoDB entries. Adding the criteria of 95% homology selected 210 entries which are new representations of the genes in EpoDB.

New feature:
We have added MEME which detects motifs conserved between sequences to the tools available on the Gene Landmarks and Gene Features Result pages. MEME was developed by Bailey, Elkan and Grundy at the UCSD Computer Science and Engineering Dept. MEME let you specify whether all input sequences have the conserved motif (or just a subset) and whether more than one motif may be found. Also available is MAST which will search EpoDB for instances of the motifs discovered by MEME.

C. Stoeckert 10/2/98

• 9/9/98: A bug has crept in and killed the Transfac query. It is off-line for now.
• 6/22/98: A page of examples is provided to give users a better sense as to what EpoDB is good for.
• 5/31/98: The BLAST Search query now uses the bioWidget BLASTView application to view results of the query. The BLASTView application provides an interactive graphical view of HSPs and allows selected HSPs to be aligned by an integrated CLUSTAL program.
  WARNING: To try BLASTView you need a browser which supports the JDK1.1 awt package such as Netscape 4.04j2. Details of the BLASTView application can be obtained at the CBIL BLASTView Demo page.
• 5/12/98: A new query has been added that returns transcription factor binding sites for a given gene name based on Transfac entries.
• 4/6/98: The "gene family" controlled vocabulary has been expanded to include developmental-stage information for the globin genes. These new terms can be used in the gene family, the feature, and the landmark queries.
• 1/26/98: The landmark query now has start of introns as a landmark to choose.
• 12/8/97: The promoter box query now allows two boxes to be searched for on the same sequence.

Server:
• We are now running EpoDB on a dedicated server, "Daphne" (the bloody one in Greek mythology). It's a Sun UltraSPARC with two 200 MHz processors. A speed-up in processing queries should be noticeable.

Looks:
• Use of frames. We eliminated the header and footer frames to make more screen space available for query pages and results. The query frame on the side is kept and made less obtrusive.

Features:
• All queries that return sequence now permit saving sequence from all or selected entries to a file in FASTA format. In addition, returned sequences from all or selected entries can be aligned using CLUSTAL.
• The latest bioWidget map and seqeunce applets are linked to the features returned by ID queries for all entries.
• For direct comparison, both EpoDB features and GenBank features are provided on the same Result page. EpoDB features are derived from GenBank after checking for syntax and semantic errors. If possible features are generated for EpoDB from existing features resulting in a richer description of the entry. Sometimes the GenBank features are so confusing that features are lost in EpoDB.
• Where available, transcription factor binding sites from TRANSFAC are added to features returned by ID query. These have links to TESS.

Expanded Queries:
• The Gene Name query now returns a table with links to protein data derived from Swiss-prot entries integrated into EpoDB.
• The Landmark query will return any available sequence for the range given. Previously, if there was insufficient sequence over the entire specified range, no sequence was returned.
• The Text Search query allows you to search protein (Swiss-Prot-derived) entries in EpoDB as well as DNA (GenBank-derived) entries.

• 5/27/97: A new way to access EpoDB entries was added: Find Entries by Text Search. A Boolean search of an outline of EpoDB entries is provided.
  A page on erythropoiesis has been added showing red blood cells at different stages of differentiation.
  Also a mistake was found in naming the HOX 1A reference gene, it should be HOX A1 (or 1F).
• 4/18/97: User-entered sequence can be used to BLAST search against entries in EpoDB. All EpoDB entries can be used or the search restricted to globin or non-globin entries. The ref library provides a non-redundant (although currently small) set of genomic entries.
• 4/18/97: The link to TESS has been upgraded to use the Filtered String Site Search. You can now limit transcription factors (TFs) used in the search by entering text to match fields in the Transfac database such as organism or cell specificity (Note:not all TFs expressed in erythroid cells are listed as such in Transfac).
• 4/14/97: About 2300 new EpoDB entries added. 64 new reference gene entries. 42 new gene names added.
  The keywords used to search GenBank for these entries are available as a link on the home page.
• 10/23/96: The link to TESS has been upgraded. TESS now uses version 3.0 of Transfac. Other new features in TESS can be found here.
• 8/5/96: Entries in EpoDB were updated from GenBank. Band 3 protein and Duffy antigen/chemokine receptor are now available for queries.
• 7/10/96: Prototype java applets displaying gene expression data and gene regulatory sites have been added which are accessible from the home page. removed from version 2.1 because the real thing is coming soon!

Looks:
• Use of frames. We're trying out frames to provide quick access to the different query pages, schema and controlled vocabulary descriptions, and links to CBIL.
• User-friendly Result pages. The submitted query is listed on the Results page in an easy-to-read form. In most cases, the results are returned as a table with clickable entries.

Features:
• Use of Java applets to display transcription unit structure and sequence. We're taking advantage of the new Java language to dynamically link gene maps and sequence. Written by Shan Dong (CBIL), the map applet allows pan/ zoom and has clickable features that highlight associated sequence in the sequence applet. From the sequence applet, highlighted sequence can be used to query EpoDB for the associated gene feature (given as a text description).
• Direct export of retrieved sequences to TESS, a transcription element search program. TESS uses the Transfac database for the searches.
• Description of gene features in GenBank format. Those familiar with GenBank should find our EpoDB descriptions of gene features easier to follow.

Expanded queries:
• Landmark query expanded to retrieve any specified sequence. Previously, a range of sequence could be specified relative only to the start of transcription. Now, other gene landmarks can ber used such as the start of translation or the start of each exon.
• New query to return types of sequence. In the "feature" query, ranges of sequences do not have to be specified. A menu is given to allow choice of features such as 5'UTR, exons, and introns.
• Promoter box query allows user to define promoter box. Previously, only a small set of choices were available for this search. Now, the user can enter and search with any desired motif.

  From Robert Graves's The Greek Myths: 1 page 80 in the
  chapter Apollo's Nature and Deeds we have the comment:


  

  His pursuit of Daphne the Mountain-Nymph, daughter of the
  river Peneius, and priestess of Mother Earth, refers apparently to
  the Hellenic capture of Tempe, where the goddess Daphoene
  ('bloody one') was worshipped by a college of orgiastic
  laurel-chewing Maenads (see 46.2 and 51.2).  After suppressing the
  college - Plutarch's account suggests that the priestess fled to
  Crete, where the Moon-goddess was called Pasiphaë (see 88.e) -
  Apollo took over the laurel which, afterwards, only the Pythoness
  might chew.  Daphoene will have been mare-headed at Tempe, as at
  Phigalia (see 16.5); Leucippus (`white horse') was the sacred king
  of the local horse cult, annually torn in pieces by the wild women,
  who bathed after his murder to purify themselves, not before (see
  22.1 and 150.1).
  

  Also from Graves's, Daphne is the daughter of Teiresias, the
  blind seer, and is taken to Apollo after the razing of her original
  temple.