Chapter 1

Bibliographic databases
Barton W. Trawick and
Johanna R. McEntyre
National Center for Biotechnology Information, National Library of
Medicine, National Institutes of Health, Bethesda,
MD 20892, USA.

Use of the literature is fundamental to the pursuit of all knowledge. Through

searching and reading, we learn what our peers are doing, develop a broader per-
spective on our field of interest, get ideas, and confirm our discoveries. During
the course of twentieth century science, ‘the literature’ has become an expand-
ing knowledge base that represents the collective archive of the work carried
out by the international scholarly community. Recent technological advances
make an increasing proportion of the literature available electronically (see
Figure 1). This chapter offers an introductory guide for molecular biologists to
stable bibliographic resources that are available over the Internet.

1 General introduction
The term ‘bibliographic databases’ has traditionally referred to the ‘abstracting
and indexing services’ for the scholarly literature. These services focused on col-
lecting the citation information and abstracts of research articles and making
them searchable. Abstracts have been the focus for the creation of bibliographic
databases because they summarize the full research article, are small enough to
re-key (the only way to capture the information before electronic publishing),
store, and search.
However, technological advances over the past decade have expanded the hori-
zons of bibliographic database creation from using abstracts only to using longer
pieces of text. Furthermore, the rise in use of the Internet has provided the
opportunity to build online, searchable literature databases that are accessible to
anyone with an Internet connection.
In response to this opportunity, publishers, libraries, and other information
providers have adopted new electronic publishing technologies to develop many
forms of online content. These include databases of journal abstracts, full-text

“chap01” — 2003/9/16 — page 1 — #1

BARTON W. TRAWICK AND JOHANNA R. MCENTYRE

100000

90000

80000

70000
Number of articles

60000

50000

40000

30000

20000

10000

0
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
Year
Total number of genetics Records with links Records with links
records in PubMed to fulltext to FREE fulltext

Figure 1 The number of PubMed records that are classified under the general MESH
term “Genetics” has grown from around 48,000 in 1992 to 93,000 by the end of 2002.
Around 1995, a number of science, technical, and medical (STM) journals began to
establish an online presence.a Since that time, the proportion of records that provide
links to online full-text articles has increased. In 2001 and 2002, around 87% of
“Genetics” records have links to online full-text articles; about 25% of these are freely
available. (The smaller proportion of free-access articles in 2002 is indicative of the
common practice of publishers to delay free access for a period of time after
publication.)
a Hitchcock S, et al. 1996. A survey of STM online journals 1990–95: the calm before
the storm.
http://journals.ecs.soton.ac.uk/survey/survey.html

articles, and books, as well as ‘internet-only content’ in the form of news and
summaries.
For the purposes of this chapter, ‘bibliographic databases’ will be consid-
ered as ‘any large, stable, collection of primarily text-based information that
is available over the Internet’. The chapter will therefore not discuss individ-
ual journal titles (although it could be argued that the online collection of
articles from a single journal constitutes a small database), nor will it discuss
the more popular health and medicine websites (though there are many to
choose from). Further, molecular biology sequence databases frequently have

“chap01” — 2003/9/16 — page 2 — #2

 BIBLIOGRAPHIC DATABASES

some literary or descriptive component; but if their focus is on data rather

than text, it will not be discussed here. As electronic publishing is in such a
state of flux, the discussion will be limited to only the most stable resources on
the web.
Many of these databases require a personal subscription, a library subscription,
or a site license, but several of the resources discussed are free to use.
This chapter is divided into three sections, based on the following types of
bibliographic information described:

1 Abstracts: bibliographic databases that contain the abstracts of journal articles

plus the citation information (e.g. author names and affiliations, the journal
title, volume, and page numbers).
2 Full-text articles: there are now several resources available on the web that offer
free access to the complete articles from life science journals.
3 Books and text-rich websites: some publishers—both traditional and new ones—
are now experimenting with the online publication of textbooks, as well as
new, information-rich websites.

2 Abstracts
When investigating a new topic area or seeking an update on a known research
area, searching an online collection of abstracts of journal articles is often the
first approach.
The strategy used to search abstracts databases is central to how successful
you will be in finding what you are looking for (or discovering things you did
not know you were looking for!). A search query that is too broad will pick up so
many abstracts as to be useless, while one too specific might be too limiting for an
expansive search. It takes practice to find the right balance and may require the
use of Boolean search constructs and techniques such as delimiting the search by
restricting it to specific fields, for example, searching only author names, or only
article titles. A general introduction to the use of advanced search techniques is
summarized in Box 1.

Box 1 Tips and tricks for searching bibliographic databases

Boolean searching
Boolean expressions (named in honour of the English mathematician George Boole)
allow the user to combine ‘AND’, ‘OR’, and ‘NOT’ operators with specific search terms
in order to create a more defined query. In most search engines, operators may be
combined (solved in order from left to right), and parentheses ‘( )’ may be employed
to clarify terms, group them together, and change the order in which expressions are
solved.

“chap01” — 2003/9/16 — page 3 — #3

BARTON W. TRAWICK AND JOHANNA R. MCENTYRE

Box 1 (Continued)
The following expressions: Will return references that contain:
Watson AND Crick at least both terms ‘Watson & Crick’
Watson OR Crick Either ‘Watson’ or ‘Crick’, or ‘Watson &
Crick’
Watson NOT Crick ‘Watson’ but not ‘Crick’
Wilkins AND Watson AND Crick at least ‘Wilkins & Watson & Crick’
(Watson AND Crick) OR Wilkins ‘Watson & Crick’ or ‘Watson, Crick & Wilkins’
or ‘Wilkins’
Wilkins AND (Watson OR Crick) ‘Wilkins & Watson’ or ‘Wilkins & Crick’ or
‘Wilkins, Watson, & Crick’
Watson AND Crick NOT Wilkins at least both terms ‘Watson & Crick’ with no
occurrences of ‘Wilkins’

Limiting searches to fields

In addition to using Boolean expressions to define a more specific output, it is also
possible to limit individual terms to fields. For example, the term ‘Crick’ may be limited
to ‘Author Name’ and ‘Nature’ may be limited to ‘Journal Name’. The number of possible
fields that are in a given database may vary, but at a minimum usually include: Author
Name, Author Affiliation, Journal Name, Article Title, Publication Date, Page Number,
Issue, and Volume.
History functions
The results of two or more searches can be combined to form a third output, or additional
terms may be added to results from previous searches through the use of ‘history’ func-
tions. This can be particularly useful for reducing large search results into smaller, more
focused ones or for combining several different terms with a single common term. For
example, independent queries for ‘cancer’, ‘DNA repair’, ‘1995’, ‘Vogelstein’, ‘human’,
and ‘mouse’ could be used in various permutations and combinations (linked by Boolean
expressions) to form new queries. Additionally, some bibliographic databases can be
customized so that useful queries can be stored for future use.
For an example of how these search techniques can be combined to search PubMed,
see Protocol 1.

2.1 Databases—in all their forms

There are several abstracting and indexing services available online, many of
which require a subscription. There is considerable content overlap among the
major bibliographic databases, and for this reason your library is unlikely to
subscribe to all of them.
When considering the use of any of these databases, it is important to make
a distinction between the database itself (i.e. the physical collection of abstracts)
and the access route into the information. Several of the large databases can
be accessed from more than one place, because the owners of the data (i.e. the
abstracts collection) lease or sell their data, or have allowed service providers to
furnish a portal to the information.
Table 1 lists the major abstracts databases for molecular biology, along with
the owner of the database and a list of access points into the data. MEDLINE, for
example, is one of the most widely used abstracts databases. Some of the abstracts
of MEDLINE can be distributed freely (those under copyright require permission
to reproduce them), so many organizations have developed clones or interfaces
to MEDLINE that generate alternative portals to the same information. This can

“chap01” — 2003/9/16 — page 4 — #4

 Table 1 Abstracts databases

Resource Produced by Examples of access Free access* URLs

PubMed/MEDLINE The National Library of PubMed Yes http://www.pubmed.gov
Medicine (NLM) BioMedNet Yes http://research.bmn.com/medline
Ovid No http://www.ovid.com/
BIDS Yes http://www.bids.ac.uk/
ISI Citation Database Institute for Scientific Web of Science No http://www.isinet.com/isi/journals/
(Web of Science) Information (ISI)
Current Contents® Institute for Scientific Current Contents Connect No http://www.isinet.com/isi/journals/
Information (ISI) Ovid No http://www.ovid.com/
BIOSIS Previews® BIOSIS BIOSIS No http://www.biosis.org/
(comprising biological abstracts Ovid No http://www.ovid.com/
and biological abstracts/RMM®
Pascal Institut de l’Information BIDS Yes http://www.bids.ac.uk/
Scientifique et Technique
EMBASE Elsevier Science EMBASE.com No http://www.embase.com/
Ovid No http://www.ovid.com/
The Cochrane Reviews The Cochrane Library The Cochrane Library Yes http://www.update-software.com/
(abstracts) abstracts/crgindex.htm

“chap01” — 2003/9/16 — page 5 — #5

* In cases where access to the database is not free, consult your library for subscription information.

5
BARTON W. TRAWICK AND JOHANNA R. MCENTYRE

provide a useful addition to a publisher’s website, or it may produce an interface

in a language other than English.

2.1.1 The databases

2.1.1.1 PubMed/MEDLINE
PubMed was developed at the National Center for Biotechnology Information
(NCBI), within the National Library of Medicine (NLM), USA. It encompasses the
over 12 million abstracts in MEDLINE, and currently covers about 4000 biomedical
journals, dating back to 1966. MEDLINE abstracts have a controlled vocabulary
associated with them known as Medical Subject Heading (MeSH) terms. Several
terms are assigned to each MEDLINE abstract, and are used for indexing articles
to provide a consistent way to retrieve information.
As well as enabling abstract searches (e.g. see Protocol 1), PubMed offers the
following additional functions:

1 Links to biological sequence information, including data such as GenBank

protein and nucleotide sequences, and macromolecular structures.
2 Links to the full-text of journal articles (about 4000 journals are currently
linked in this way). Whether the full text can be viewed without purchasing the
journal depends on the journal policy (see section below on full-text articles).
3 Links to ‘Related articles’. For each abstract, similar articles in the database
have been identified, based on a statistical analysis of words and phrases found
in the abstract text. This is an easy way to expand on a PubMed search when
a useful abstract has been found.
4 Links to resources outside of the NLM. The ‘LinkOut’ feature allows other
providers of information, such as organism-specific databases like FlyBase,
to link to related abstracts.
5 Links to textbooks. A new collaborative project at the NCBI is linking the
content of textbooks to PubMed abstracts to serve as background information
(see Section 4.2).

PubMed is primarily a biomedical database that historically has not collected

abstracts from non-medical areas of molecular biology. However, more recently,
the scope of PubMed has widened to include coverage of those areas, such as the
plant sciences. PubMed does have the significant advantage that it can be used
free-of-charge from anywhere in the world.

Protocol 1
Using PubMed
The PubMed page (Figure 2(a)) consists of the following: (1) A sidebar that contains links
to PubMed information and services; (2) A query box for entering search terms; (3) A
feature bar that contains links for advanced searching; and (4) Links to other integrated
molecular biology databases.

“chap01” — 2003/9/16 — page 6 — #6

 Sidebar Tool bar Query box Integrated database links

A
OMIM
link

C
D

Figure 2 The PubMed search interface. (a) The PubMed page has a sidebar that links to
related services and information, links to integrated databases, a query box for entering
search terms, and a tool bar that contains links for advanced searching. (b) Previous
searches can be viewed using the history feature. Searches can be combined through
use of search numbers and Boolean operators. (c) The ‘Limits’ feature allows searches
to be constrained to various information fields, such as Author Name or Review Article.
(d) The field restrictions are found only in the ‘All Fields’ pull-down menu.

“chap01” — 2003/9/16 — page 7 — #7

BARTON W. TRAWICK AND JOHANNA R. MCENTYRE

Protocol 1 continued
As an example of how a PubMed search can be conducted, we will look for review arti-
cles on CD95 (or Fas, a lymphocyte receptor) and apoptosis, written by P. H. Krammer.

Method
1 Enter the search term ‘apoptosis’ in the query box and click the ‘Go’ button to the
right of the box to initiate the search (Figure 2(a)). Conducting a search using a broad
term without any field restrictions usually returns a large number of hits; in this case,
more than 70,000 citations are found.
2 PubMed retains the most recent search term in the query box. Click the ‘Clear’ button
to the right of the query box to remove the previous search term (‘apoptosis’) and
replace it with the new search term ‘CD95’ and click ‘Go’. The search for ‘CD95’
returns several thousand references.
3 Click on the ‘History’ tab located in the feature bar. The results of the two previous
searches are now displayed chronologically in a numbered list. These results may
be reviewed individually by clicking on the number of returns for each query in the
‘Results’ column (Figure 2(b)).
4 Searches stored in History may be combined using Boolean operators (see Box 1) to
form a new search. Search for references that contain both ‘apoptosis’ and ‘CD95’ by
typing ‘#1 AND #2’ in the query box and clicking on the ‘Preview’ button. Selecting
‘Preview’ will display the search results in History summary format, rather than
listing each article found. (Note: PubMed is case-insensitive for search terms but
case-sensitive for Boolean operators: make sure ‘AND’ is in capitals.)
5 New queries may be combined with previous searches. To find references associ-
ated witha the name P. H. Krammer, type ‘#3 AND krammer ph’ into the query box
(Figure 2(b)) and click ‘Go’.
6 Queries can be limited to various fields such as: journal name, author name, title
word, MESH term, publications type, publication date (or date range), or language.
To employ limits in PubMed, click on the ‘Limits’ tab in the feature bar. Limit the
‘Publication Types’ pull-down menu to ‘Review’ (Figure 2(c)). Check that the search
term ‘krammer ph’ is still present in the query box and click ‘Go’. The result displays
a summary view of all review articles associated witha P. H. Krammer that contain
the query terms ‘apoptosis’ and ‘CD95’.
7 Limits will remain in effect for subsequent searches unless they are deselected by
clicking the check box to the left of the ‘Limits’ tab in the feature bar. Additional limits
are located in the ‘All Fields’ pull-down menu (Figure 2(d)). For a more comprehensive
account on advanced searching of PubMed, consult the help documentation, listed
in the sidebar on the PubMed homepage.

a When searching for authors by last name only, rather than for abstracts that merely

cite the author’s last name, the search should be carried out with the field limited to

“chap01” — 2003/9/16 — page 8 — #8

 BIBLIOGRAPHIC DATABASES

Protocol 1 continued
Author Name. For example, a search for ‘Crick’ without any field limits will return
abstracts that contain the word ‘Crick’, such as in the term ‘Watson–Crick base pair’.
Similarly, a search for articles in the journal ‘Cell’ needs to be executed with the field
limited to Journal Name, otherwise the results will list any abstract that contains the
word ‘cell’.

2.1.1.2 Web of science

The Institute for Scientific Information (ISI) produces the ‘Web of Science’—an
interface to the ISI Citation Database that contains more than 5300 scientific
articles, dating from 1980, that is updated weekly. The Web of Science is a
subscription-based service, available from many (but not all) university libraries.
The Web of Science shares some features with PubMed, such as links to biological
sequence information and full-text articles, but also has some that are unique:

1 Links to related articles. The way in which related articles are calculated in the
Web of Science differs from the related articles of PubMed. In Web of Science,
the list of records related to a given article consists of papers that cite at least
one source also listed in the original (parent) article, with the source that has
most common citations listed first.
2 Links to (i) the Derwent Innovations Index, a patent database; (ii) BIOSIS Pre-
views, a database of references to primary journal literature, meetings, and
books; (iii) ISI Chemistry Server, for newly reported structural chemistry.

For many molecular biologists, one of the most valuable attributes of the Web
of Science comes from the use of the citations associated with each abstract.
Through the references cited within an article, it is possible to:

(a) View the abstracts of all articles cited in the original (parent) article,
(b) Find all articles published, since the original (parent) article, that have cited
it, and
(c) Find all the articles that have cited a particular author.

2.1.1.3 Current contents

The Web of Science also interfaces with the ISI Current Contents databases, for
which a subscription is required. Current Contents used to be a paper publication,
distributed weekly and consisting of the contents of recently published journals,
divided into broad subject categories, such as the Life Sciences (coverage of about
1400 journals). The Current Contents database can be searched, abstracts of arti-
cles found can be viewed, and from there the table of contents of the journal
issue can be displayed and browsed.

“chap01” — 2003/9/16 — page 9 — #9

BARTON W. TRAWICK AND JOHANNA R. MCENTYRE

2.1.1.4 EMBASE
EMBASE (1974–present) is a bibliographic database produced by Elsevier that
covers over 4000 journals in the biomedical and pharmacological sciences. Its
online presence now incorporates selected MEDLINE records, thus increasing the
scope and scale of EMBASE to over 13 million abstracts. Like PubMed and Web
of Science, EMBASE has links from appropriate abstracts to selected full-text arti-
cles and gene sequence information. EMBASE is available by library subscription
only.

2.1.1.5 The Cochrane Abstracts

The Cochrane Reviews is a collection of reports that collate and summarize pub-
lished health care evidence on a wide range of medical disorders and conditions.
The target audience is very broad, ranging from those receiving care, to those
responsible for research, teaching, funding, and administration of health care at
all levels.
The reports are written and maintained by international panels of clinicians,
who are organized into groups on the basis of area of expertise. There are cur-
rently about 50 Collaborative Review Groups that cover areas such as breast
cancer, schizophrenia, HIV/AIDS, and tobacco addiction. Once a review is written,
it is checked regularly and updated as needed.
While a subscription is required to access the full Cochrane Reviews, anyone
can browse or search the Cochrane Abstracts without charge. The abstracts alone
are quite substantial (usually about 300–500 words). They outline the background
for the study, the source data, search strategy, and criteria for inclusion, and then
state the results and conclusions. The Cochrane Abstracts, while more focused on
clinical trials and therapies than basic molecular biology, are a high-quality and
useful adjunct for those who work on molecular biology problems with clinical
applications.

2.1.1.6 BIOSIS Previews

BIOSIS Previews is made up of two databases: Biological Abstracts, which con-
tains about 12 million records from more than 5000 journals, and Biological
Abstracts/RRM, which covers reports, reviews, and meetings—information not
formally published in scientific research journals. This includes references to
items from meetings, symposia, and workshops, review articles, books, book
chapters, software, and US patents related to the life sciences. It covers the biolog-
ical sciences, from biochemistry to zoology, and is available by subscription only.

2.1.2 Access providers: BIDS and Ovid

BIDS and Ovid are companies that aggregate databases created by other orga-
nizations into convenient packages for libraries to use. BIDS may be the
best-known bibliographic service for academics in the United Kingdom and Ire-
land. It provides access to a number of databases, some of which are freely

10

“chap01” — 2003/9/16 — page 10 — #10

 BIBLIOGRAPHIC DATABASES

available, and links to full text articles via Ingenta Journals (see Section 3).
Many databases and services formerly provided by BIDS, including Medline, are
now provided free to UK academics via the ISI’s Web of Knowledge interface
(http://wok.mimas.ac.uk).
Your library may also use Ovid as a provider of several bibliographic databases,
including BIOSIS Previews®, Current Contents®, EMBASE (Excerpta Medica
Database), and MEDLINE, among others.
Database aggregators often implement databases in their own way, so the
interface for searching the databases may have several features that differ from
the implementations of other providers.

3 Full text of research articles

Most of the databases described above concern abstracts of published research
articles. Although not considered traditionally as bibliographic information, no
discussion of online text resources would be complete without considering the
increasing availability of full-text articles.
Several thousand molecular biology journals are now available in electronic
form (Figure 1)—most are online counterparts to paper journals, but some are
online-only publications (see Box 2 for a summary of the advantages of online
articles over articles printed on paper). All can be viewed via a web browser,
providing that, with a few exceptions such as the Journal of Clinical Investigation, you
have a subscription. However, more recently, some journals have made articles
from back issues freely available, and new publishing ventures that offer free
access to articles are emerging (see Table 2).

Box 2 Advantages of online journals over paper journals

Searchable content
Articles in digital format may be searched for words and phrases. Most bibliographic
databases provide a search engine that allows for content matching across all entries.
Once an article has been obtained, the ‘Find’ feature in your web browser can be used
to search within the article for specific words and phrases.
Hypertext links
Online articles displayed in HTML can exploit hypertext linking to create connections
between related content. Links can be made from references cited in the text, to its
listing in the bibliography, or to external information sources such as PubMed abstracts,
referenced citations, errata, sequence information, macromolecular structures (PDB
files), or even the author’s home page.
Multimedia
The content of traditional printed journals is restricted to what can be presented on
paper. However, online journals are able to ‘add value’ to articles with movies, audio, and
the inclusion of large data sets (an entire genome sequence, for instance). Additionally,
use of color figures does not generally represent a higher publishing cost for online
journals as it does for print journals.

11

“chap01” — 2003/9/16 — page 11 — #11

BARTON W. TRAWICK AND JOHANNA R. MCENTYRE

Box 2 (Continued)
Accessibility
Electronic articles can be accessed over the Internet rather than visiting a library. This
is particularly useful for those in remote locations. Downloaded electronic articles can
be stored on your personal computer.
Flexible publishing model
Some scientific journals make their content available online before the printed copy.
Some journals even provide a ‘rolling model’ of publication where articles are accessible
online as soon as they are accepted for publication. Manuscript submission, online peer
review, and access of electronic content may be provided by some journals through the
Internet.

Table 2 Online full-text journals

Resource Produced by No. of Free URL

journalsa accessb
Science Elsevier 1100 No http://www.sciencedirect.com/
Direct Science
Link Springer-Verlag 500 No http://link.springer.de/
Interscience Wiley 300 No http://www.interscience.wiley.com/
BioMed Current 130 Yes http://www.biomedcentral.com/
Central Science
Society and Highwire Press 340 Some http://highwire.stanford.edu/
small and individual journal
publisher URLs
online
journalsc
PubMed The National 150 Yes http://www.pubmedcentral.gov
Centrald Library of
Medicine
a Journal figures given in round numbers. Figure represents the total number of journals in each resource;
not all of these may be life science journals.
b In cases where access to the database is not free, consult your library for subscription information.
c HighWire Press enables small publishers to make their journal content available online. It is not the
publisher of these journals.
d PubMed Central is an active archive for journal content; it is not a publisher.

3.1 Access to the full text of research articles

In the absence of a search engine that indexes a good proportion of full-text life
science journals, the best route to finding full-text articles is not always obvious.
As mentioned above, access to an article is only possible if you or your library
has a subscription, or if the article is made freely available. Below we outline the
most common and useful routes to online journal articles.

3.1.1 Access through abstracts databases

Most of the databases listed in the previous section can make links between
abstracts and the corresponding online full-text article. There will be a link that

12

“chap01” — 2003/9/16 — page 12 — #12

 BIBLIOGRAPHIC DATABASES

leads the user seamlessly to the article if the following is true:

1 The journal (more specifically, the journal issue) is published online.

2 The publisher of the journal has agreed with the database to make the article
available via this route.
3 You or your library subscribes to the journal, or the publisher makes the article
freely available.

For example, a search of an abstracts database will result in a list of ‘hits’

consisting of the citation information for each article retrieved by the query (see
Box 1). If the abstract satisfies points (1) and (2) above, then there will be a link to
the journal publisher’s website (this may only become apparent when viewing
the complete abstract rather than the citation information). Clicking on this link
will take you to the full-text of the article if point (3) is satisfied. Many of the
freely available articles can be found in this way, and, as an example of scope of
access, about 4000 journals currently have links from PubMed abstracts to their
respective articles on the publisher’s site.

3.1.2 Access from publisher sites

Many publishers do not collaborate with all bibliographic databases to allow
access to their journals, and the most conservative may only allow access to their
journals by logging-on directly to their own website. In these cases, the only way
to access the full-text is through your library’s interface to the journal, or by a
direct visit to the journals’ website, if you hold a personal subscription. Here we
will list some of the most significant places where there is a collection of full-text
articles (see also Table 2).

3.1.2.1 HighWire Press

HighWire Press works with scientific societies and publishers to create online
counterparts to their print journals. There are currently over 340 journals that
are available at HighWire, of which about 150 now offer free access to back issues
of the journal.
The period of time after which the article becomes freely available depends on
the policy of the journal. Some journals, such as British Medical Journal (BMJ) have
an immediate free-access policy (i.e. anyone can look at the most current version
of BMJ). However, most HighWire journals operate under a delayed-release policy
for free full-text articles, ranging from 2 months to 5 years, with most opting for a
1–2 year delay. In total, there are now (Spring, 2001) around 250,000 free articles
available. HighWire allows a basic search across all the journals they collaborate
with, although the free articles are not clearly delineated.

3.1.2.2 Individual publishers

Many publishers have developed their own online interfaces to their journal
databases. Some of the largest of these are listed in Table 2, although there are

13

“chap01” — 2003/9/16 — page 13 — #13

 BARTON W. TRAWICK AND JOHANNA R. MCENTYRE

many smaller collections. For all these sites there is usually some free intro-
ductory content, but the journal content is almost always available only on a
subscription basis.

3.1.3 Archives for full-text articles

Publishing journals online is still a relatively new enterprise. Now that there is
a substantial volume of information available over the Internet, the question
of how to effectively archive the data and make the best use of the electronic
medium for searching and linking becomes obvious.
A recent initiative called PubMed Central, based at the National Library of
Medicine USA, is aimed towards creating an archive for full-text life science jour-
nal articles that can be browsed and cross-searched freely. The idea is that any
journal article available via the PubMed Central site can be viewed by anyone
with an Internet connection from anywhere in the world.
Currently, about 150 journals are making their content available via PubMed
Central, and though small at present, the potential of this kind of initiative for
the future makes PubMed Central worthy of mention.

4 Books and text-rich websites

While books have been less evident than journals in making the transition from
paper to electronic form, a few online texts do exist, although most require
a subscription or site license. A growing trend is for books to have associated
websites for further information and corrections (as this book has∗ ). These are
usually listed prominently in the book. Furthermore, as biological content on the
Internet evolves, so do content-rich websites that do not fit into any traditional
bibliographic mold; this category of bibliographic resource is not well-defined, so
here we will discuss just two of the larger and more stable resources (see Table 3
for URL).

4.1 Online Mendelian inheritance in man

Online Mendelian Inheritance in Man (OMIM) is a catalogue of human genes and
genetic disorders (see Table 3 for URL). It now contains about 15,000 records, and is
authored and edited by Dr Victor A. McKusick and his colleagues at Johns Hopkins
and elsewhere. The online version has been developed by the NCBI.
The OMIM database is usually searched using the name of a genetic disorder
or the name of a gene to retrieve records, and it is possible to use Boolean search
constructs as well as field limitations, such as chromosome number (see Box 1).

Table 3 Online books

Resource URL
Online Mendelian http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM
Inheritance in Man
Online books at NCBI http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Books

14

“chap01” — 2003/9/16 — page 14 — #14

 BIBLIOGRAPHIC DATABASES

OMIM does not contain any figures or graphics, but it does have links to PubMed,
gene, and protein information. OMIM is now one of the databases integrated
with PubMed, and can be accessed for searching by clicking the OMIM link on
the PubMed search page (see Figure 2).

4.2 Online books

A project to put biomedical textbooks online, make them searchable, and inte-
grate them with PubMed and other data resources has recently begun at the
NCBI (see Table 3 for URL). There are currently about 24 books participating in the
project, which broadly cover the subject areas of basic molecular and cell biology
and genetics (Figure 3); more books are set to become available in the near future.
The book collection may be searched directly, using a similar interface to
PubMed. In addition, all PubMed abstracts have a ‘Books’ link; clicking on this
link brings up a facsimile of the abstract with hyperlinked terms and phrases that
lead to the most relevant sections of the book(s) for the linked phrase. PubMed
abstracts are rich in information, but they do not explain the terms or concepts
used, so linking abstracts to books as background information may help address
this shortfall. The quantity and subject area of hyperlinked phrases in an abstract
will depend on how much the content of the abstract overlaps with that of the
books available.
While the complete contents of the book are free to use in this way, for some
books it is not possible to navigate across the whole book content, from chapter
to chapter. In these cases, access is limited to ‘stand-alone’ chapters or sections.

4.3 Text-rich websites: a word of caution

Any web search engine can also be used to search for molecular biology infor-
mation. Many publishers, biotech companies, research labs, teachers, and others
display information that can be browsed freely.
Information found in this way should be carefully evaluated. Be aware that
anyone can publish almost anything on the Internet, so a key factor in assessing
the validity of any information found is the reliability of its source. It is important
to assess what qualifies the individual or organization to publish the information,
and what their motivation for doing so has been. As with any literature search,
the information found should be cross-checked and critically evaluated before
believing.

5 Summary
Bibliographic information on the Internet for molecular biologists continues to
grow. This chapter must really be considered a snapshot, serving as an introduc-
tion to the potential for exploring online literature resources. For this reason we
have chosen to discuss only the most stable of resources, and have not discussed
the specific use of any one search interface. The websites and databases discussed
undergo constant evolution, and new resources are continually launched and
developed. The Internet moves faster than the print world; we hope that this
chapter will at least be in the same race for some time!
15

“chap01” — 2003/9/16 — page 15 — #15

BARTON W. TRAWICK AND JOHANNA R. MCENTYRE

Figure 3 How to access the books at NCBI. (a) All books can be searched directly from
the books homepage (see Table 3 for URL), as well as indirectly, through hyperlinked
phrases in PubMed abstracts. Each PubMed abstract obtained by searching PubMed has
a ‘Books’ link. Clicking on this link displays the same abstract with some hyperlinked
phrases, as shown here. (b) Executing a books search or clicking on a hyperlink within a
PubMed abstract displays a summary list of books in which that term is found. The
number on the right indicates the number of book sections that are relevant for the
term. This link leads to a book-specific list of sections, figures, and tables. Figures and
tables are indicated by the icons shown. (When less than 20 relevant book sections are
found, the book summary step is omitted.) (c) The section, table or figure titles lead to
the book content. The books are displayed as one chapter section per page, and it is
possible to navigate around a minimum of one chapter at a time. The books contain links
to the figures and tables of the book, PubMed abstracts, and in the future will be more
extensively linked to molecular biology information.

Acknowledgements
We would like to thank Kathi Canese and Edwin Sequeira for carefully reading
this manuscript.

16

“chap01” — 2003/9/16 — page 16 — #16