There are two types of things you can use EpoDB for.

• Look up information.
• Do sequence analysis.

• bring together information about genes from different sources.
• create a set of reference genes that we have complete and verified information on.
• can return the same specific region of sequence from a number of related genes for further analysis.
• provide visualization tools that makes it easy and efficient to interpret the information in EpoDB or from the analyses.

  Looking up information:

  The way to look up information about a gene is to use the queries under the heading
  Find entries in EpoDB based on:

  Example 1: Getting GenBank type information on genes.

  
  Suppose you wanted to get information on the erythropoietin receptor.
  • Go to the Gene Name page. You can get there from the Gene Name link in the frame to the right.
  • Click on erythropoietin receptor in the menu box.
  • Click the "Execute EpoDB query" button. You should get a Result Table in about a minute or less.
  • Click on the "36239" link in the "EntryID" column.
  
  Ta Dah! You should get a Result of EpoDB ID Query page. If you scroll through you'll see:
  • A bunch of information and features in a GenBank style format. After the the first bit of stuff that tells you things like accession number and def (definition line) you get a column of features like "exon" followed by cryptic numbers and gibberish (/parents?). The numbers tell you where (Location) on the sequence the feature is located.
  • More information and features in GenBank format. These are the actual GenBank features.
  • t_factor_bind_site. These are transcription factor binding sites obtained from Transfac.
  
  Note you can get these directly through the links at the top.
  You can also see most of this information in our spiffy bioWidget viewers if you follow the "graphical form" link. Warning this can take a long time (several minutes) to load up so you might want to check your gel or take a bio-break now.

  Questions you might have:

  The gene I'm interested in isn't on the the list to choose from. Does that mean it's not in this database?
  Not necessarily, it means that we don't have a "reference" version of your gene. We may have an mRNA version or a gene fragment but we don't have a complete version. To check our whole listing of genes (and proteins) use the "Entries by text search" query (available on the frame to the right). We're working on a way to improve this so that all entries in EpoDB can be retrieved using a name for a reference gene.

  So the Result table gave me the mouse erythropoietin receptor. Is there any way I can get the human?
  The answer to this is pretty much the same answer for the previous question. Try the "text search" query. Type "erythropoietin receptor" (including the quotes) and you'll get 17 DNA entries or 3 protein entries which include the human gene.

  What's with all the id numbers and computerese? Can't you just give us the relevant information?
  Sorry. The next version of EpoDB will get rid of all that stuff. It helps us keep track of whether things are working right but you don't need to see it.

  I'll have to take your word that the graphical version is spiffy because it never worked on my computer. Any suggestions?
  Yes, use Netscape 4.05 on a PC or UNIX station. The problem is that the programming language we use to make the graphical views (Java) needs the latest version for it to work right - the Mac version only sort of works.

  Why are you showing the GenBank information twice?
  The two versions (EpoDB's and the original GenBank) are not the same. For example, the GenBank version does not come with 5'UTR and 3'UTR features. You have to figure them out which is what we did with a computer program. So our version gives you the COMPLETE annotated gene. The GenBank version is provided so you can see what we based our version on. Sometimes it's an opportunity to show off because we can deduce a lot from minimal GenBank data. Othertimes, our program doesn't work (like on the non reference genes) so you can at least get the GenBank info.

  Do sequence analysis:

  There are 3 ways to do sequence analysis using EpoDB. One is is to use the queries under the heading
  Retrieval of sequence for gene and gene family entries as specified by:
  another is to use the BLAST page,
  and the third way is to use our promoter-box query under the heading
  Search gene sequences for:
  

  Example 2: Retrieval of sequence for gene and gene family entries as specified by Landmark.

  
  Suppose you were interested in comparing the promoters of gamma-globin genes for insights as to why these genes are expressed during the embryonic stage in some animals but in other animals (like humans) the genes are expressed during the fetal stage of development. You can use the gene landmark query to retrieve promoter sequence for gamma-globin genes in the two classes (embryonic and fetal expression) and then align them using CLUSTAL to see where areas of sequences conservation occur.

  First we'll get the proximal promoters of fetally-expressed gamma-globin genes, then get the promoters for embryonically-expressed gamma-globin genes, and then aligh them together.

  • go the "Sequence by Landmark" query (use the link in the frame to the right).
  • click the gene family radio button.
  • click on "fetal globin, beta-like" in the menu under gene family.
  • enter "-200" for start of range and "1" for end of range.
  • choose "5' mRNA boundary" for point chosen.
  • click on Execute EpoDB Query.

  Ta Dah! You should get a Result of EpoDB Query Result page with the information you specified listed at the top.
  This is followed by a table with 10 entries; the first two are sheep and goat followed by G-gamma and A-gamma globin promoters for chimpanzee (pan), gorilla, orangutan (pongo), and humans. The gamma-globin gene was duplicated and recruited for fetal expression in old world apes. The beta-like globin genes in ungulates (sheep, goats, cows) went through a different type of duplication. In fact the whole cluster (epsilon, gamma, eta, and beta) was triplicated and one of the beta gene descendants was recruited for fetal expression.

  For our analysis let's just focus on the apes to keep things simple and focussed on the gamma-globin gene.

  • Click on the boxes next to entries 3-10 to select them.
  • Check that the "save" button at the bottom of the table is selected.
  • Click "Submit."
  You should have the 8 promoter sequences you selected in FASTA format.
  Copy these (or save them as a file) and we'll add them to a list of embryonic promoters which we'll get next.
  • go the "Sequence by Landmark" query (use the link in the frame to the right).
  • click the gene family radio button.
  • click on "embryonic globin, beta-like" in the menu under gene family.
  • enter "-200" for start of range and "1" for end of range.
  • choose "5' mRNA boundary" for point chosen.
  • click on Execute EpoDB Query.

  Now you've got a table with 37 entries! As before, we will just focus on the gamma-globin genes.
  Note that not all the entries are 200 bp long even though that's what was requested. The EpoDB query returns whatever it can in the range specified so if only 126 bp are available for lemur (#14) that's what you get.

  • Click on the boxes next to entries 11, 13, 14, 35, and 37 to select them.
  • This time click the "multiple sequence alignment" button at the bottom of the table is selected.
  • Click "Submit."

  OK. Now you have the embryonic promoters for gamma-globin genes (a rabbit, a lemur, and some monkeys) in a box ready to be aligned. Add the fetal globin promoters to the list and hit the "perform alignment" button.

  Presto! You've got an alignment of the gamma promoters; the top 5 sequences are the embryonic followed by the 8 fetal. The "+" at the bottom indicates complete conservation which you can see around the "TATA" (actually the AATAAA box in this case) and upstream of the CCAAT box. Note that just upstream of the TATA box is a sequence "GGCGGCTGGCT" which is well conserved among the fetal 8 but not in the embryonic 5! Could this be something special like a potential transcription factor binding site?

  Let's check it out.

• go the "Sequence by Landmark" query again.
• click the gene family radio button.
• click on "fetal globin, beta-like" in the menu under gene family.
• enter "-200" for start of range and "1" for end of range.
• choose "5' mRNA boundary" for point chosen.
• click on Execute EpoDB Query.

This time instead of selecting entries to align we will use the TESS link.

• Click on the TESS link next to the sequence in entry #10 (human).

A new browser window pops up with our sequence ready to be scanned for potential transcription factor binding sites using the TESS (Transcription Element Search Software).

• Choose "Organim Species" for Factor Selection Attribute in the Element and Match Filter section.
• Type "Homo" in the Search Pattern filed to restrict our search to just human transcription factors.
• Choose to view sequence with text or Java in the output section. Java can take a very long time depending on your network connection and computer. Choose text if you are not sure.

  Wow. Of the 649 human transcription factors in the Transfac database, 416 had binding sites which could be used to search our sequence. These are listed at the bottom of that page. Let's focus on our region of interest which is around 00151.
  Hey, there's an AP-2 site (T00035) which overlaps our sequence "GGCCGGCGGCgGG." Maybe it's involved in the fetal recruitment. It's back to the lab to find out but you get the idea as to why retrieving sequence and analysing can be useful.
  Let me know if you find out anything!
  -Chris Stoecker