Transcript Apoptosis-associtated pathways are induced vy Phytophthora

Interrogating the DRASTIC
Gene Expression Database
Gary Lyon
DRASTIC
Database Resource for Analysis of Signal Transduction in Cells
www.drastic.org.uk
30 April 2004
Aim of DRASTIC
To understand signal transduction in response to plant
pathogens and other environmental stresses.
To assist with putting into context the results of our own
gene discovery work within the PPI Programme
and
Publicity !
Why do we need ‘DRASTIC’?
• Published gene expression data is not searchable.
• Too much data to remember e.g. microarray data.
• Cannot match ‘unknown’ genes with prior expression
data (14.2% of entries in the database are ‘unknown’).
• Gene names associated with certain accession numbers
change with time.
• Cell biology is complex. [Simple answers to complex problems
are always wrong]
For example
• One gene can have a variety of names :
HBZip
homeobox domain HD-zip
homeobox protein
homeobox domain zipper protein
transcription factor, homeobox protein
• Names can be wrong:
‘HB AtHB-14 like’ should be ‘AtHB-9’
‘Htf9C’ should be ‘RNA methyltransferase-related’
‘endo 1,4-beta-mannosidase like’ should be ‘protein kinase family’
• Names can be confusing:
‘HSR201 like’
‘RSH2 :Rel-SpoT homology’
www.drastic.org.uk
Access database
• Incorporates published data from microarrays and
Northerns of ESTs regulated by various treatments
(i) Environmental stress e.g. drought, NaCl, high and low
temperatures
(ii) Pathogens and elicitors (salicylic acid, ethylene,
jasmonates)
• 424 references
• 266 treatments
• 67 plant species
• 10,193 gene accessions
Selection by Gene name
treatment 1
Potential
signalling
networks
1
treatment 2
treatment 3
4
7
2
5
3
6
Davina Button
Funded by a 1 year PGRA grant from Carnegie Trust awarded to:
University of Abertay
– Dr Les Ball, Dr Louis Natanson (Computing)
– Prof Kevan Gartland, Dr Jill Gartland (Biotech.)
– Davina Button (RA)
University of Edinburgh
– Prof Peter Ghazal (GTI; Scottish Centre for Genomic Technology and
Informatics)
University of St Andrews
– Dr Ishbel Duncan (Computer Science)
Aim:
–To build an intelligent and generic system for new hypothesis
formulation from complex biochemical pathway databases.
‘Road Map’
Options with the new database
Genes induced by BTH
pathogen induced – incompatible (Arabidopsis)
Pathways e.g glycolysis enzymes
Conversion of glucose to pyruvate
• Wrong pathway
• Insufficient data
Possible interpretations:-
• Some errors (different time points? low homology!)
• Evidence of another pathway
Text mining
1. Les Ball (Abertay),
2. Prof Bonnie Webber (School of Informatics, Edinburgh
University),
3. CABI.
•
Data input and
•
Data analysis
Could be used to provide a putative relationship between
genes/proteins based on existing knowledge in the
literature. This model could be combined with
information in the gene expression database to provide a
draft version of a regulatory gene network.
Web stats - Location of users
Impact factors ?!
DRASTIC
Database Resource for Analysis of Signal Transduction in Cells
SCRI
University of Abertay
Gary Lyon
Les Ball
Adrian Newton
Louis Natason
Bruce Marshall
Alasdair Houston
www.drastic.org.uk
Can we group treatments?
Genes up-regulated by Sulphur depletion
Another example
The same gene can have different accession numbers – a big
problem with genes of unknown function.
However, by converting accession numbers into AGI
numbers we have shown that for the following ESTs
down-regulated by :chitin (viz H37231, R90140, T41806),
drought (viz AV823744),
ethylene (viz R90140),
low oxygen (At2g10940) or
sodium chloride (AV823744),
or up-regulated by salicylic acid (R90140, H37231)
are all the same gene viz At2g10940
Number of entries in the Gene expression Database - examples
up-regulated
down-regulated
Arabidopsis
potato
tomato
Nicotiana tabacum
pepper
rice
5052
168
393
258
113
234
1246
8
213
87
0
43
Treatments
ethylene
salicylic acid
jasmonates (methyl)
jasmonic acid
105
330
344
78
20
146
135
2
Pathogens
Ecc
Eca
P. infestans (incompatible)
P. infestans (compatible)
35
3
15
51
0
0
1
3
wounding
436
690
546
510
187
263
248
63
Abscisic acid
359
46
Total in database
7127
1828
Plants
Environmental cold
drought
stresses
sodium chloride
What else could we do with the data?
• Identify potato and barley orthologs of stress induced
genes
• Map the position of the stress inducible genes
• Statistical analysis of signal transduction genes
• What are the differences between different plant
tissues e.g. roots v. leaves.
Information from Maleck et al., Nature Genetics (Dec 2000) 26, 403-410
Out of 50 accession numbers checked (March 2004):• 26 (52%) were correctly identified
• 3 (6%) were wrongly identified (though 2 of these could be classed
as ‘additional information being made available’ with only 1 really
wrong.
• 13 (26%) are newly identified with a gene name (these were
originally described (‘no homology’)
• 8 (16%) remain unknown but have an AGI number (these were
originally described as ‘no homology’)