The Semantics of Genomic Analysis
- 1. The Semantics of Genomic Analysis Robert Stevens BioHealth Informatics Group School of Computer Science University of Manchester robert.stevens@manchester.ac.uk
- 2. The Collaboration “Developing a GRID-based system for integrating and exploring data from comparative genomics, to discover biological knowledge that can not be discovered from any one source” Collaborative BBSRC project – 5 sites across the UK http:// www.comparagrid.org
- 3. 3 Introduction • The general problem • An architecture for getting answers • Forming the questions and making distinctions • Some observations
- 4. Sleeping Cows: Unknown Genes 4 ??? Cow Infected with African Trypanosomiasis Cow chromosome and known Genes
- 5. Sleeping Mouse Model with QTL 5 Mouse Infected with African Trypanosomiasis Mouse QTL region on chromosome, encompassing multiple genes
- 6. We can Find Out About Mouse Resistance 6 Mouse Infected with African Trypanosomiasis Mouse QTL region on chromosome, encompassing multiple genes Genes A, B, C, D in QTL are involved in Trypanosomiasis resistance
- 7. Use Mouse to Tell us about Cow Infer presence and order of Cow genes from presence and order of Mouse genes
- 8. 8 Requirements • Ask for maps that that contain marker from species x in species y, z, etc • What chromosome or part of chromosome does a map model? • What is the gene order in species x and species y? • Also notions of similarity, homology, and, of course, synteny • Orthology and paralogy
- 9. The Fluxion Stack Raw data Raw data Syntax Semantics Aggregation Pub svc Trans svc integrator query data
- 10. 10 Query Semantics • Query OWL class interpreted as – K query – (poor man’s) epistemic closure of query – Against knowledge-base exposed by that data-source, not The World • Result is a knowledge-base – All entailed by queried KB (it’s a subset) – Can be statements in the original KB, or any statements that always follow • Contains at least the statements needed to – Allow a reasoner to classify all the individuals who match the query correctly – Preferably using properties, not asserted types (a-box preferred over t-box,don’t over-commit) – Speed for accuracy – Implementation complexity for data-volume – Return all instances of known classes e.g. db table with minimal filtering – if in doubt, return it
- 11. Role of Ontology in Fluxion • Fluxion – Uses semantics of OWL – Not any domain-specific information – Any domain • A domain ontology defines what Fluxion integrates • Developing a ‘good’ domain ontology is – Hard work – Poorly scoped – No widely-validated methodology – Biologist ¹ Modeller so language gap
- 12. ComparaGRID Ontology Scope 12 • Genetics – Markers and Maps • Genomics – Genomes and Sequences • Comparative Aspects – Evolutionary relationships, Similarities • Physical entities – Chromosome, Organism
- 13. Models and Representations • We ‘know’ things about physical biological things – We know other things about the way we render information about these things – What reconciles a map and a sequence of a thing is the thing itself – A “knowledge anchor” has representation is representation of 13 Physical Biological has model is model of Maps Entities Sequences
- 14. 14 Modelling Maps • A map is an abstract model of a physical thing – Good for ordering limited knowledge – Minimal explicit biology in a map, just implicit biology in the labelling of things – Maps can be modelled as a combination of lines and blobs: 0 A B C D E 1) Line: model-part 2) Blobs: model-parts 3) Map: Model of ordering of blobs within line
- 15. 15 Modelling ‘Markers’ • ‘Marker’ is overloaded term • Meaning here is captured by linking lines and blobs up to physical things 0 AB C D E Map is model of (ordering of things on) chromosome or region AB C D E ‘Line’ is model of chromosome or region ‘Blob’ is model of (detectable) region
- 16. Strategy: Domain and Application Ontologies 16 Upper classes Domain classes Derives Classes used by data model(s) Datatypes Informs
- 17. TAMBIS AllOver Again • …, but now we have different technologies; bio-ontologies galore; different resources. • Mainly SQL queries over RDBMS • Can actually answer the questions we ask • Rather than mapping down to the resources; we map down from the global model and up from the resources into an OWL world 17
- 18. 18 Main Points • Its hard: Sophisticated ontologies don’t appear by magic • Open world good, open world bad • Need to do local closure: A poor man’s epistemic operator… • It isn’t completely correct • A sophisticated ontology in a SW application • Highly axiomatised ontologies for the Semantic Web?
Editor's Notes
- #3: <number>
- #5: Slide 1 Sick cow on left Arrow in middle Chromosome on right Question marks above arrow Text says: (left) Cow Infected with African Trypanosomiasis (right) Cow chromosome and known Genes
- #6: Slide 2 Sleeping mouse on left Arrow in middle Chromosome on right with QTL and multiple genes highlighted Text says: (left) Mouse Infected with African Trypanosomiasis (right) Mouse QTL region on chromosome, encompassing multiple genes
- #7: Slide 3 Sleeping mouse on left Arrow in middle Chromosome on right with QTL and multiple genes highlighted Text says: (left) Mouse Infected with African Trypanosomiasis (small text) (right) Mouse QTL region on chromosome, encompassing multiple genes (small text) (centre) Genes A, B, C, D in QTL are involved in Trypanosomiasis resistance (large text)
- #8: Slide 4 Sleeping mouse on left below mouse chromosome Arrow in middle Sick cow on right below cow chromosome Text says: (centre) Infer presence and order of Cow genes from presence and order of Mouse genes
- #10: <number>
- #12: <number>