GBCH723 Bioinformatics and Genomics

Introduction

Course Goals:

To identify important databases for biomedical research
To explain methods for interfacing with databases effectively
Discussion of papers and techniques that utilize bioinformatic and genomic data

There is no required text. Here are a couple of books that I have found helpful:

Bioinformatics: A Practical Guide to the Analysis of Genes and Proteins 2nd Edition by Andreas D. Baxevanis (Editor) – a good general overview

Bioinformatics: Sequence and Genome Analysis by David W. Mount – more intense explanations of algorithms

Beginning Perl for Bioinformatics by James Tisdall - good introduction to writing your own programs for use in bioinformatics. Does not assume extensive computer knowledge.

Bioinformatics is defined as:

The use of computers in solving information problems in the life sciences, mainly, it involves the creation of extensive electronic databases on genomes, protein sequences, etc. Secondarily, it involves techniques such as the three-dimensional modeling of biomolecules and biologic systems.

by the Online Medical Dictionary.
 

Internet

Since computers, and usually, the internet, are so heavily involved in the use of bioinformatics, a brief introduction to how the internet itself works may be beneficial.  Much of this info was obtained from UNH InterOperability Lab  and  PC Lube & Tune.

Lets start by clicking on a web page link using your internet browser:
for example:

 1  129.81.133.1 (129.81.133.1)

 2  tidewater-et-4-1.net.tulane.edu (129.81.255.93)

 3  newsouth-atm-1-0-0.net.tulane.edu (129.81.255.70)

 4  abilene-houston-pos-oc3.tis.tulane.edu (129.81.255.2)

 5  atla-hstn.abilene.ucaid.edu (198.32.8.34)  University Corporation for Advanced Internet Development

 6  wash-atla.abilene.ucaid.edu (198.32.8.66)

 7  wash-abilene-oc48.maxgigapop.net (206.196.177.1)  Mid-Atlantic Crossroads (MAX)

 8  clpk-so3-1-0.maxgigapop.net (206.196.178.46)

 9  wash-nlm.maxgigapop.net (206.196.177.34)

10  130.14.38.185 (130.14.38.185)

11  micasaweb.nlm.nih.gov (130.14.22.106)

 

There is the possibility that unscrupulous people may pretend to be other computers and intercept private data, like credit card numbers. This is why some transfers use secure, encrypted transfers (https instead of http) which prevent others from deciphering what is being sent.

Important Databases:

Genbank and EMBL DNA sequence databases

Both contain virtually all known sequences, including complete genomes

Genbank and SWISSPROT protein sequence databases

Mostly translated coding sequences from the DNA database

Important file formats for both protein and DNA databases are:

GenBank: protein example - DNA example

PDB: Protein Data Bank 3-D structural database

Genome databases, most accessible through Entrez

Currently there are:
more than 100 complete Bacterial genomes
15 complete Archeael genomes
18 complete Eukaryal genomes, including Human
and hundreds of viral genomes

Last year there were:
53 complete Bacterial genomes
11 complete Archeael genomes
10 complete Eukaryal genomes
and hundreds of viral genomes

PubMed citation database

Thousands of Titles and abstracts from medically relevant journals dating back to the 1960's. Some older citations also available. Powerful searching capabilities essential for identifying articles of interest. Similar databases available for other disciplines (i.e. agricultural)

PubMed introduction and tutorial

This page is condensed from the NCBI PubMed Tutorial Pages . You may find the full tutorial quite useful.

When you enter search terms on the main PubMed search page,  the PubMed server processes your request to attempt to identify what type of search you are attempting: are you looking up an author name, journal title, subject area, or phrase from the article abstract?  It accomplishes this by filtering your search terms through successive lists to identify the types of terms you provide and use them effectively. This process is called:
Automatic Term Mapping

PubMed compares your search terms against several lists of search terms to determine what you are looking for. It checks four lists in order  and stops looking once it finds a match:
 

1. MeSH (Medical Subject Heading) Translation Table
2. Journals Translation Table
3. Phrase List
4. Author Index

The MeSH Translation Table contains:

MeSH terms and Subheadings

(searching synonyms for MeSH terms)

Chemical Names of Substances

Full journal titles

MEDLINE title abbreviations

International Standard Serial Numbers (ISSN)

The Phrase List contains several hundred thousand phrases generated from:

MeSH

Unified Medical Language System (UMLS)

Chemical Names of Substances

Author Searching

The format for author searching is last name plus initials.
PubMed will automatically truncate the author's name to account for varying initials.

If the term is not found, PubMed will then search the individual words in All Fields.

You can also try putting a phrase in double quotes if the results returned are not what you expected. This will force PubMed to look for the words as a phrase, but it bypasses the Automatic Term Mapping, so you might want to try doing some searches both with and without double quotes.

Truncation

You can truncate a word with the asterisk (*) wildcard This will causes PubMed to return all matches that begin with the truncated string of text. (e.g. enzym* will match enzyme, enzymes, enzymology, enzymatic, etc.) Truncation also turns off Automatic Term Mapping, so the results will be different than nontruncated searches.

Stopwords

PubMed also refers to a list of commonly found words that are referred to as "stopwords ." these are very common words which would match almost every citation and so they are skipped.

The list of stopwords is from PubMed's Help Page.

Stopwords

 

a	did	it	perhaps	these
about	do	its	quite	they
again	does	itself	rather	this
all	done	just	really	those
almost	due	kg	regarding	through
also	during	km	seem	thus
although	each	made	seen	to
always	either	mainly	several	upon
among	enough	make	should	use
an	especially	may	show	used
and	etc.	mg	showed	using
another	for	might	shown	various
any	found	ml	shows	very
are	from	mm	significantly	was
as	further	most	since	we
at	had	mostly	so	were
be	has	must	sum	what
because	have	nearly	such	when
been	having	neither	than	which
before	here	no	that	while
being	how	nor	the	with
between	however	obtained	their	within
both	I	of	theirs	without
but	if	often	them	would
by	in	on	then
can	into	our	there
could	is	overall	therefore

Operators

You can use Boolean operators (AND, OR, NOT) to direct your search. These must be entered in UPPERCASE. Operators are processed left-to-right unless you use parentheses to specify the order.

Once you click the "Go" button. Your search is performed and the first 20 hits are displayed in a Summary format:

Author name(s):

Title of the article:

Brackets indicate a title translated from a foreign language.

Source: a brief journal citation.

Identification number: A PubMed Unique Identifier (PMID) is included on each record.

Links: Includes links to Related Articles and databases, when available.

You can easily scan this first page of citations and see how many of them are really related to what you were trying to find. Though only the first 20 citations are displayed by default (in reverse chronological order) you can see how many total articles matched your search. If you got a surprisingly small or large number of hits, or if there seem to be a high percentage of extraneous hits, you might want to click on the "Details" button in the upper gray box.

Details Button

Clicking Details displays:

 

The PubMed query box shows exactly how PubMed performed your search using the Automatic Term Mapping. It may have found a synonym in the MeSH headings and used that instead of one of your original terms.

You can edit the search used and run the edited search by clicking "Search".

If the search worked really well, you can save it as a web link by clicking "URL" This formats your search as a URL link your web browser can save as a bookmark to repeat the search at a later date. You can also use the "Cubby" system described below.

The "Result "section shows how many hits you got, and links you back to your hits. The translations section describes how each term of your search was interpreted.

The database is PubMed, and The User Query is what you typed in to begin with.

You can select Publication types (like reviews) from another menu. You can limit searches to specific dates or trials involving subjects in specific age groups, gender, or human/non-human.

You can require that hits have Abstracts, though some reviews do not have abstracts, nor do articles indexed before 1975.

You can have even more control over limits by using the Preview/Index Feature. You can add search terms by limiting to specific fields, but you can preview the number of results by clicking on the preview button.

By clicking on index, you can also look up search terms in the index (for example the index of MeSH terms). Items can be added to the search window using the AND, OR, or NOT buttons.

Different searches can be combined using their Query number found in the Preview/Index page, a more extensive list is found on the History page. (ex, #4 AND #5). Note that these query numbers disappear after 1 hour of inactivity, so you can't use yesterday's Query number tomorrow and get the same result.

You also cannot use these numbers to save your results as a URL in the details window, but you can manually cut and paste the query lines together to save them.

Now that you have constructed the perfect search, you can select the perfect format for displaying results. The default is 20 summary results, but you can choose another format: Other available formats for citation display can be chosen by selecting from the list of choices listed under "Summary":

 
Brief format includes:

• First author
• First thirty characters of the title.
• PMID #
• Links

Abstract format provides the summary information in addition to:

• First Author affiliation
• Abstract, if one is present.
• Links to full-text of the article at provider's Web site, if available.
• Links to Related Articles, Books, LinkOut, and databases.

Citation format is similar to abstract, but also includes:

• MeSH terms.
• Chemical Names of Substances, if any are present.
• Grant numbers, if any are present.

MEDLINE format is a text file with identifying letters before each field. It is most useful for importing into bibliography programs like EndNote and ProCite.

Selecting Citations and Display Format

You can select a subset of the hits to display by clicking the box before each item. If you don't click any boxes, then all are displayed.

• Or you can click on the individual links to see the abstract format for a given citation.

Add to Clipboard

You can select individual citations to save in a clipboard on the server. This is not the clipboard on your computer. After selecting items by clicking their checkbox, click on the "Add to clipboard" link.

• The color of item numbers of the hits changes when added to the clipboard.
• If you did not click any boxes, the entire search gets loaded to the clipboard (up to the limit of 500 hits).
• You can view the clipboard by clicking the "Clipboard" link in the features bar. The Clipboard disappears after one hour of inactivity.

Save Button

You can save citations to a file on your computer by clicking the "Save" link. There is a limit of 10,000 hits. To save selected citations, pick a display format and press "Save". You will be prompted for where to save the downloaded file.

Text Button

You can have the selected items displayed as plain text by clicking the "Text" button. This may be useful for printing if your browser doesn't print the hypertext files well.

Cubby

If you set up a "Cubby", you can save your favorite searches indefinitely on the PubMed server. You have to get a username and password. You can then save your search and rerun it at a later date. Or you can run the search for new articles published since the last time you searched.

LinkOut Preferences

The LinkOut service enables publishers, libraries, biological databases, sequence centers, and other Web resources to display links to their sites on records in PubMed.

You can use Cubby to set which links are displayed by

• Adding icons to the Abstract and Citation formats
• Hiding providers from the LinkOut format

When you are logged into Cubby, PubMed displays LinkOut providers according to your preferences.
Related Articles - Compares words from the title, abstract, and MeSH headings to identify articles similar to the selected article.

Related Articles

NCBI Databases

These are the NCBI databases that may be linked to from individual PubMed citations:

• Protein: Protein sequences from SWISSPROT, Protein Information Resource (PIR), Protein Research Foundation (PRF), Protein Data Bank (PDB), and translated protein sequences from the DNA sequences database.
• Nucleotide: DNA sequences from GenBank , European Molecular Biology Laboratory (EMBL), and DNA Data Bank of Japan (DDBJ).
• PopSet: Sequences submitted as a set from a population studies.
• Structure: experimentally-determined, three-dimensional structures.
• Genome: Records and graphic displays of genomes.
• Taxonomy: Index of organisms represented in the sequence databases.
• OMIM: A catalog of human genes and genetic disorders.
• Books provides links to terms described in selected molecular biology textbooks.

 

GBCH723 Home Page