Transcript Slides_for_iPC_Presentation, 15Oct2009

Overview of Bio-Array Resource DataViz Tools
• Framework
• Webservices
• Expression Browser
• eFP Browser
• ePlant
BAR Framework
• The Bio-Array Resource consists of numerous MySQL DBs. There is also
associated code written in Perl, Python, Ruby, and some Java for exploring
gene expression data, protein-protein interactions, subcellular localization,
orthologs, and protein structure
• Expression data are from in-house or public DBs, such as GEO or the PDB,
and are stored in a MIAME-compliant format. Details on experimental setup,
genetic background (whether the material is from a mutant plant), etc. are
captured.
In general, we try to capture such details using standardized ontologies,
such a the Boyes system for describing the age of plants, GO etc.
Unfortunately not all of these ontologies are suitable for our needs (e.g.
Plant Ontology for now). MinSeq may be suitable for NGS data…
• The BAR runs off on a Dell webserver with 2 Dual Core Xeon Processors
(2.33GHz), 4 GB of RAM, and ~80 GB of redundant disk storage.
Webservices
• Many older BAR tools use direct SQL queries within the code to retrieve
data
• More recently, we have implemented BioMoby based webservices to e.g.
return eFP Browser images to remote clients
• We are also finding it convenient to implement lighter weight JSON
(Javascript Object Notation) webservices for internal use – this makes
development a lot faster and as a bonus provides others with direct,
computer-readable access to our data (expression data, protein-protein
interaction data, orthologs etc.)
The Bio-Array Resource: e-Northerns w. Expression Browser
Experiment Identifier
Experiment Category
Growth Stage darker=older
Tissue type e.g. green=leaf
Control/Treatment Flag
black=treatment, white=control
Mutant/WT Flag
white=WT, blue=mutant
Timecourse darker=later
Cluster Tree
co-expressed genes are grouped
Functional Classification
’Barcode’ indicates GO category
AGI IDs, My Gene Categories
colouring as specified by user
Expression Levels
red=higher expression level
Toufighi et al. (2005), Plant J 43(1): 153-163
Protein Interactions if present
Looking at DHS3 response with the eFP Browser
Mode allows viewing in Absolute,
Relative, and Compare modes
Signal Threshold to set a
maximum for colour scale
Data Source to choose different
AtGenExpress projects and other
projects, e.g. Birnbaum set.
Expression Level distribution
shows how maximum level
compares with all other genes,
and maximum level in any set
Clickable tissues hyperlink to
the NASCArrays, GEO or
literature record for the sample
Expression Level scale
red=higher expression level
Winter et al. (2007) PLoS One 2(8): e718
Browser
array_element_name locus
(from affy_ATH1_array_elements-2006-01-06.txt at TAIR)
261585_at AT1G01010
261568_at AT1G01030
261584_at AT1G01040
261579_at AT1G01050
261569_at AT1G01060
261576_at AT1G01070
261577_at AT1G01080
261583_at AT1G01090
261578_at AT1G01100
261580_at AT1G01110
261570_at AT1G01120
261575_at AT1G01130
AtGenExpress
data
At-TAX
data from
Tuebingen w.
webservices
New for 2008 – cross-species expression viewing
New for 2009 –
ePl
: seamless viewing of km- to nm-scale data
BAR.utoronto.ca/ePlant
David Di Biase, Geoff Fucile, Shokoufeh Kohdabandeh, Hardeep Nahal
3DDI.org
NGS mRNA-seq data could be displayed with BAR tools