MetNet 3: Searching and visualizing plant metabolic pathways

1
MetNet 3 Searching and visualizing plant
metabolic pathways
Yves Sucaet1,2, Eve Syrkin Wurtele1
1) Genetics, Developmental Cellular Biology,
Iowa State University, Ames, IA USA
http//www.gdcb.iastate.edu 2) Bioinformatics and
Computational Biology Interdepartmental program,
Iowa State University, Ames, IA USA
http//bcb.iastate.edu
Abstract
Interaction with the portal
Future plans
MetNet 3 is a novel online platform to retrieve
information on plant metabolic and regulatory
networks from MetNetDB. Pathways and subnetworks
visualized with MetNet3 represent user-selected
data types, including information flow from genes
to metabolites, interactions, and feedback loops
that induce (post-)transcriptional perturbations.
Availability http//www.metnetdb.org/metnet3
Although the observational data collected in
surveillance efforts describe annual mosquito
population dynamics, heterogeneity is inherent in
this dataset because measurements took place in
different locations according to voluntary
participation by county agencies. Challenges and
solutions that evolved as a result of working
with and displaying these heterogeneous data are
presented below.
Implementation of a formal API A relational
database and web interface has been created and
functions both for entering observational data,
and tracking information through the years in
spatiotemporal terms for each mosquito species
observed. Data format standardization Offer data
to outside sources in standard formats (currently
only SBML is supported) so that MetNet data can
be analyzed using third-party software (such as
Cytoscape, E-Cell, CellDesigner etc) Offer
webservices The conversion to standard data
formats in addition to the availability of a
SOAP-based webservice API leads to a new
interaction model between the different MetNet
applications that is wider than was previously
the case
Introduction
Biological network databases provide a platform
for integration and viewing of combined
high-throughput transcriptomic, proteomic and
metabolomic data, literature evidence, focused
experimental results, and annotation. In
addition, such databases can be designed to
enable computational analysis of experimental
data in the context of the known and hypothesized
biological network. These analyses enable the
development of experimentally testable biological
hypotheses about the function of genes, proteins
and metabolites, as well as potential metabolic
and regulatory interactions. MetNet
3 is a web-portal to MetNetDB, a biological
network database for Arabidopsis and other plant
species. The database is designed to encompass
and integrate metabolic and regulatory
interaction networks. The MetNetDB is contained
in a labeled graph model that facilitates
integration of annotation data with known or
hypothetical biological interrelationships.
Biological entities and interactions are
represented as nodes. Biological properties of
the entities (these include synonyms, entity
type, subcellular location, reference, gene
annotation and metabolite formula) and
interactions (these include interactiontype,
reversibility and EC number for enzymes) are
stored as corresponding node labels. Edge
labels store coefficients and kinetic
parameters for the interactions. The new
web-portal supports the ability to visualize and
query the networks, as well as compose custom
personalized networks.
Conclusion
MetNet A team effort
Iowa-mosquito.net presents an easy-to-use and
effective means to manage and analyze sizable and
metadata rich population datasets. Furthermore,
the database and web-interface have been designed
with portability and extensibility in mind
(available for windows as well as UNIX/Linux).
The web-pages have been separated into multiple
subcategories according to their functions.
Additional functions can be incorporated quickly
on a modular architecture. Currently, the
database only contains New Jersey Light Trap
(NJLT) data. Additional datasets are available
over the same extensive time-period for other
trap types. Different traps serve different
purposes, and certain species may be easier to
catch than others with specific methods. Thus,
adding data from the other traps may present an
interesting comparison regarding the efficiency
of different trapping methods among the different
mosquito species and regions.
Acknowledgements
The mosquito surveillance program at Iowa State
University was spearheaded and run by Dr. Wayne
Rowley until his retirement in 2005. His efforts
led to the bulk of the data represented herein.
Surveillance efforts are funded by the Iowa
Department of Public Health and supported by the
Agriculture and Home Economics Experiment Station
(Ames, IA) project 5111 through the Hatch Act and
State of Iowa funds.