EBI is an Outstation of the European Molecular Biology Laboratory. Avazeh Ghanbarian Paul Kersey Alessandro Vullo EBI Microme Annotation Meeting June 2011.

EBI is an Outstation of the European Molecular Biology Laboratory.

Avazeh Ghanbarian

Paul Kersey

Alessandro Vullo

EBI

Microme Annotation Meeting June 2011

[email protected] Microme Annotation Meeting June 201120.04.232

Microme: current status

• We’re behind • But we have some material:

• Curated data from the iAF1260 model (Anne’s presentation)• Databases built on EcoCyc and MetaCyc • Integration of different data sets• A working reaction/pathway projection process

• Currently based on Ensembl Compara to identify orthologous proteins• This is replaceable for future projections if a better system

is proposed• Generic rules for inferring complex/pathway completeness


Microme: data decisions

• Integration work is not finished• EBI has brought 3 databases to this meeting

• DB1: EcoCyc + MetaCyc• DB2: EcoCyc + MetaCyc + curated iAF1260 data• DB3: EcoCyc + projections

• Initial projection to ~30 species targetted for Biolog data• We can use all 3 databases to assist the curation process

• Need to decide the basis for going forward – what is our reference database? Which database are we going to curate into?


Microme: outputs

• Sets of inferred reactions for use in model building• Semi-automated feedback loop for using model-building

experience to improve reference data• A UI to offer Microme reference and non-reference data to users


Microme: web portal

• Before launch, EBI will prepare a lightweight portal giving access to all tools promoted by partners

• EBI will provide two components to this:• A genome browser for the Microme reference species (using

Ensembl technology)• This will also provide access to the Compara-based

orthology results• An application for viewing pathways (using Reactome

technology)• We have been integrating the source data into a

Reactome-style DB• There is a curation tool for the direct addition of new

pathways into this database


The Reactome world view

• Reactome has a top down (i.e. pathway-centric) view of reactions in a species

• Data model and display can store reference pathways and projections • Projections dynamically recalculated at intervals

• Reactome’s default is to project from one species to all others• Microme will (in future) project from one generic reaction bin

(containing all reference reactions) to all specific species• Reactome’s data model is based on a recursive chain of “events”: in

the data model, a reaction is simply an event in one or more pathways• Curator UI reflects the data model• User UI uses pathway as organising paradigm

• “is_a” and “part_of” relationships in Event hierarchy


Microme interface: current status

• At this meeting, we only have interfaces based on the old version of the Reactome code

• The new code is much nicer and more powerful• Interactive SBGN-based interface

• Allows upload of user expression data etc.• Allows visualisation of projected pathways against the backdrop

of the reference (shows what is present/absent against the backdrop of the full pathway).

• This requires us to draw pathway diagrams for our pathways• This is a manageable task for the E. coli pathways

• 300 pathways, many can be drawn in a few minutes using the pathway curation tool


Reactome UI


Zoom

Pathway dictionary

This event

Pathway diagram

You are here


Pathway diagrams can be complex (EcoCyc pathways less so)

About this object


External links

Internal links


External links

Internal links

Inferred entities in other species


Reactome data may be of less use to us than Reactome software


External links

Internal links




Analsysis tools


User interface: conclusions

• This interface will work for us with the present data model• But need to draw pathway diagrams first

• Mostly (much) simpler than Reactome pathways


This workshop

• Understand how the curation tool works• Curate some new reference pathways from literature• Preliminary assessment of projection data

• Is it any good?• How could it be improved to aid curation?

• How could it be used?• Potential for writing analysis scripts to explore gaps, etc.

• How can we integrate this with Biolog data• No formal time for this, but please talk to me if interested


Microme: booting the data

• To use either version of the Reactome interface, we should organise data as pathways

• We have fitted the reactions curated by SIB and CEA into the EcoCyc framework of pathways• If all the reactions in a pathway are not curated, using just

some reactions might lead to the pathway becoming disconnected


Microme: using MetaCyc

• We have also tried integrating MetaCyc• Using only experimentally supported data

• Issues• Many non-bacterial pathways• Reference pathways and complexes may have exemplars

from different species • Don’t necessarily have a consistent set of exemplars to

define a reference pathway• Taxonomical resolution is various• Some of the external curated Cyc databases not-usefully

described in MetaCyc data file• Too many pathways for us to draw


MetaCyc: conlusions

• MetaCyc is fit for its own purpose, but it may make us ill trying to swallow it

• start with EcoCyc, and potentially progressively add additional sets of experimentally approved reactions from certain additional species


Pathway comparison in current UI





UniProt entry for the catalyst in this species

EBI is an Outstation of the European Molecular Biology Laboratory. Avazeh Ghanbarian Paul Kersey Alessandro Vullo EBI Microme Annotation Meeting June 2011.

Documents

microme annotation meeting

microme interface

microme reference species

nonreference data

reference pathways

curated data

reference reactions

source data