Web-based tools for Bioinformatics; A (free) introduction to (freely available) NCBI, MUSC and World-wide Bioinformatics Resources.

When and Where---Wednesdays at 1pm Room 438 Library Admin Building Beginning September 10.

Overview September 10

We are going to look at the National Center for Biotechnology Information (NCBI) main page to review something of the scope of the NCBI. Here is the NCBI site map. NCBI maintains the GenBank archive, coordinates daily with the DDBJ and EMBL counterparts to exchange newly submitted data. NCBI has information hubs which function as archive only but has developed a growing list of curated databases the aim of which is enhanced and more precise bioinformation retrieval.

Keeping up with the growth in database size is worth reexamining from time to time. Here's one page that will link to most statistical summaries. Current notes about Genbank release may be found at this site.

There are a number of on-line materials and tutorials which NCBI has prepared which can help you orient yourselves in this Bioinformation Age. Its a good idea to review this Science primer for your sense of how this bioinfomatics world is organized. There is even a "field guide" to the NCBI resources. Check Here for an NCBI problem set to work through. There are a number of books now on-line at NCBI On-line Books . The European Bioinformatics Institute also offers an excellent site for bootstrapping your knowledge base on bioinformatics. Here is the EBI 2can site. 2can has explanations and tutorials and much more.

Introduction/Scope

The central software of the NCBI is the ENTREZ system. Here is an on-line help document for ENTREZ and a specific ENTREZ tutorial.

The concept of "neighbors" is one of the most valuable features of the ENTREZ system. These neighbors are provided via the "Related" records link in Entrez search results. These neighbors are precomputed by NCBI in a database relative manner. The "neighboring methods" are listed below. The value added by this feature is that you can often locate information that otherwise might not appear in a straightforward search. ENTREZ serves as a central information hub for the following databases.

The ENTREZ basic search is simply to type in a phrase and see what happens. You might get lucky and you will get related resources from which you might relatively quickly get your answer. There are however much more sophisticated methods of searching. The limits, preview-index and Cubby links can aid you construct more sophisticated searches. The History mode allows you to combine searches. The Details mode allows you to view how the system has interpreted your query; this is a handy way to learn the syntax for more sophisticated query building. Cubby in particular is useful if you want to regularly scan ENTREZ for new examples of the same types of references.

The ENTREZ page at NCBI shows the main integrated databases of the Resource.

Practical Use of ENTREZ

In this worked example we will start with the general search in the PubMed database for the literature referring to "caspase". We will then focus the search using various means and finally use the system to expand a search. The use of these tricks should apply to any of the databases of the ENTREZ system. In subsequent weeks we will use the other databases. Today we will stick mostly with PubMed.

Open the ENTREZ PubMed link and type the word "caspase" .

On September 4, 2003 at 12:45pm OVID search retrieved 10,583 references. ENTREZ Pubmed 11,349 (up from 11,182 Aug 27), SciFinder Scholar 23,305 containing the term and 23,344 with the "concept"

Limit the Caspase search by requiring that caspase appear in the title ( not just anywhere in the searchable text) eg "caspase[titl]"

This reduces the number of hits to 3000 or so.

It may prove interesting to see a concise view of the search progress. The "Preview/Index"
menu allows you to add terms in simple or more complex ways to the current query. The resultant
line shows you the form of the query and the number of resulting hits.

Next lets use the boolean operator "AND" to require that the word "apoptosis" also appear in our search. eg "caspase[titl] AND apoptosis"

This drops the number of hits further but just down to around 2700 articles.

Now lets amend the search and try a more drastic limiting procedure. Lets search for caspase in the title, and the word ischemia. Lets use the "Limits" menu to search for articles referring to human and then limit by date as well to articles published Jan 1 2003 to Aug 27, 2003.

This results in just one publication. So, by judicious use of various limiting features we have reduced the initial 11000 plus references to just one. We have severely narrowed the search.

Note that by selecting the "DETAILS" menu we can see (and edit) the actual interpretation of the query terms as well as see what the results might be in summary terms. By selecting the "HISTORY" menu you can see previous searches; each indicated by a "#" sign and a number. You can then "combine" searches in the history listing by typing say "#2 AND #6" or "#1 NOT #2" etc.

Now we can employ the useful concept of neighboring. This will find us similar articles. We are now broadening the search. Click on the "Related Articles" link to the far right of the citation.

This finds us 95 articles.

Looking through the list we see regular "text" icons and a bit further down, another icon.

Clicking on the J Biol Chem article we can see the abstract and a link to the full text of the article.

If we next click on the "Links" item for this single entry we can see a protein entry. Clicking that in turn brings us to a protein sequence entry for a "scramblase". From there we can enter the sequence and possibly the structure databases.

These exercises serve to illustrate the basic features of the ENTREZ system. You can start with rather general word and locate may articles in PubMed which in one or another way refer to the word. This hit list can be extensivly reduced by carefully applying limits. From any single reference, we will be able to use the "neighboring" functions to retrieve related articles;thereby expanding a given search. In some circumstances, these references may have links to sequence and/or structural information. This is why the BCR page places the ENTREZ link under the "Relational Databases" link. That is the tremendous power of the ENTREZ system to combine neighboring and relational features and allow the searcher to customize searches to locate articles.

Note two things. Not all journals are indexed by PubMed. PubMed is not instantaneous. Some articles are indexed quickly. Others take a while. Watch the National Media (NPR/PBS, NYTimes etc) for references to new research findings and publications and then watch to see how long it takes PubMed to list the citation.

As researchers, you will find that publication can be a bit late in the game. Attendance at national/international meetings and direct communication can still keep you more or less up-to-date in a given field. These Internet tools are superb to aid you getting up to speed on a complex research topic.

 

Sample Questions/Data

Try some of the same searches with other databases eg Nucleotide or Protein divisions of ENTREZ. You will find that the text searching concepts can be applied to these sequence databases as well. The Limits menu options will however be more database specific. Be sure to examine the Links for any related entries from the other databases. Note as well that the format of the sequences as well as the format of the literature hits can be presented in various formats. The retrieval format you choose will depend on the end use of the information. Programs such as ENDNOTE require the "MEDLARS" format of a literature reference, whereas many sequence analysis programs look for the "FASTA" format. In other circumstances you might prefer "XML" format. Keep your end-use in mind.

Examples of how to keep up:

Ovid AutoAlerts, Entrez Cubby

Subscribe to a newsletter such as HumMolGen. Another option is to subscribe to Science Magazine's Editor's Choice e-mail service.

Harvest Sequence Information@EMBL
Bioinformatic Harvester (c)

Search for  

Here's the complete link to the Bioinformatics Harvester

Histone Code -nonDNA regulation of gene expression.

Cancer susceptibility OGG1

NCBI and ENTREZ do not have all the answers. Here's an example.


Suppose you wish to identify potential transcription factor binding sites in your DNA sequence. Ok bioinformatics sleuth what would YOU do? Here are some suggestions.

From a MolGen Alert you extract the following quote "the lipid leukotriene B4 (LTB4) is a major mediator of the migration of 'effector' T cells.". What's leukotriene?

 

Created by ESH 8-18-2003; updated 9-4-2003 11:30

email to Starr about this page