COMPUTATIONAL TOOLS 81
support a wider variety of queries and data types, but may be slower to gain adoption. Another effort to
accomplish agreement on data and metadata standards is the National Biological Information Initiative
(NBII), a program of the U.S. Geological Survey’s Center for Biological Informatics.
Agreement on standard terminology and data labeling would accomplish little if the data sources
were unknown. The most significant challenge in creating large-scale ecological information is the
integration and federation of the potentially vast number of relevant databases. The Global Biodiversity
Information Facility (GBIF)^66 is an attempt to offer a single-query interface to cooperating data provid-
ers; in December of 2004, it consisted of 95 providers totaling many tens of millions of individual
records. GBIF accomplishes this query access through the use of data standards (such as the Darwin
Core) and Web services, an information technology (IT) industry standard way of requesting informa-
tion from servers in a platform-independent fashion. A similar international effort is found at the
Clearinghouse Mechanism (CHM),^67 an instrumentality of the Convention on Biodiversity. The CHM is
intended as a way for information on biodiversity to be shared among signatory states and made
available as a way to monitor compliance and as a tool for policy.
Globally integrated ecological databases are still in embryonic form, but as more data become
digitized and made available by the Internet in standard fashions, their value will increase. Integration
with phylogenetic and molecular databases will add to their value as research tools, in both the ecologi-
cal and the evolutionary fields.
4.3 Data Presentation,
4.3.1 Graphical Interfaces,
Biological processes can take place over a vast array of spatial scales, from the nanoscale inhabited
by individual molecules, to the everyday, meter-sized human world. They can take place over an even
vaster range of time scales, from the nanosecond gyrations of a folding protein molecule to the seven
decade (or so) span of a human life—and far beyond, if evolutionary time is included. They also can be
considered at many levels of organization, from the straightforward realm of chemical interaction to the
abstract realm of, say, signal transduction and information processing.
Much of 21st century biology must deal with these processes at every level and at every scale,
resulting in data of high dimensionality. Thus, the need arises for systems that can offer vivid and easily
understood visual metaphors to display the information at each level, showing the appropriate amount
of detail. (Such a display would be analogous to, say, a circuit diagram, with its widely recognized icons
for diodes, transistors, and other such components.) A key element of such systems is easily understood
metaphors that present signals containing multiple colors over time on more than one axis. As an
empirical matter, these metaphors are hard to find. Indeed, the problem of finding a visually (or
intellectually!) optimal display layout for high-dimensional data is arguably combinatorially hard,
because in the absence of a well-developed theory of display, it requires exploring every possible
combination of data in a multitude of arrangements.
The system would likewise offer easy and intuitive ways to navigate between levels, so that the user
could drill down to get more detail or pop up to higher abstractions as needed. Also, it would offer good
ways to visualize the dynamical behavior of the system over time—whatever the appropriate time scale
might be. Current-generation visualization systems such as those associated with BioSPICE^68 and
Cytoscape^69 are a good beginning—but, as their developers themselves are the first to admit, only a
beginning.
(^66) See http://www.gbif.org/.
(^67) See http://www.biodiv.org/chm/default.aspx.
(^68) See http://biospice.lbl.gov/home.html.
(^69) See http://www.cytoscape.org/.