1. Alan Ruttenberg School of Dental Medicine Applications Alan Ruttenberg Oral Diagnostic Sciences Clinical and Translational Data Exchange

2. Alan Ruttenberg School of Dental Medicine Goals of this section Present a sampling of semantic web applications – OBO Annotation – Immunology 2

3. Alan Ruttenberg School of Dental Medicine Depictive ontology, patient-centered queries

4. Alan Ruttenberg School of Dental Medicine Investigating the connection between features of histocompatibility proteins and disease 4 http://neurocommons.org/page/HLA_Structure_Variation

5. Alan Ruttenberg School of Dental Medicine Variations in this allele select distinct ?hla_res where { ro:has_part ?hla_res; sc:blast_match ?allele. ?hla_res sc:reference_sequence_item ?item. ?allele sc:perturbation ?ares. ?ares sc:reference_sequence_item ?item. } SPARQL query results drive visualization 5 Contact points in this HLA chain select distinct ?hla_res where { ro:has_part ?hla_res. ?hla_res sc:chain_contact ?peptide_res}

6. Alan Ruttenberg School of Dental Medicine Looking for Alzheimer disease targets Signal transduction pathways are considered to be rich in “druggable” targets - proteins that might respond to chemical therapy CA1 Pyramidal Neurons are known to be particularly damaged in Alzheimer’s disease. Casting a wide net, can we find candidate genes known to be involved in signal transduction and active in Pyramidal Neurons?

7. Alan Ruttenberg School of Dental Medicine A SPARQL query for processes involved in pyramidal neurons prefix go: prefix rdfs: prefix owl: prefix mesh: prefix sc: prefix ro: select ?gene_name ?process_name where { ?pubmed_record ?p mesh:D017966. ?article sc:identified_by_pmid ?pubmed_record. ?gene_record sc:describes_gene_or_gene_product_mentioned_by ?article. ?protein rdfs:subClassOf ?restriction. ?restriction owl:onProperty ro:has_function. ?restriction owl:someValuesFrom ?restriction2. ?restriction2 owl:onProperty ro:realized_as. ?restriction2 owl:someValuesFrom ?process. {{?process ro:part_of go:GO_0007166} union {?process rdfs:subClassOf go:GO_0007166 }} ?protein rdfs:subClassOf ?parent. ?parent owl:equivalentClass ?restriction3. ?restriction3 owl:onProperty sc:is_protein_gene_product_of_dna_described_by. ?restriction3 owl:hasValue ?gene_record. ?gene_record rdfs:label ?gene_name. ?process rdfs:label ?processname. } Mesh: Pyramidal Neurons Pubmed: Journal Articles Entrez Gene: Genes GO: Signal Transduction Inference required

8. Alan Ruttenberg School of Dental Medicine Results DRD1, 1812adenylate cyclase activation ADRB2, 154adenylate cyclase activation ADRB2, 154arrestin mediated desensitization of G-protein coupled receptor protein signaling pathway DRD1IP, 50632dopamine receptor signaling pathway DRD1, 1812dopamine receptor, adenylate cyclase activating pathway DRD2, 1813dopamine receptor, adenylate cyclase inhibiting pathway GRM7, 2917G-protein coupled receptor protein signaling pathway GNG3, 2785G-protein coupled receptor protein signaling pathway GNG12, 55970G-protein coupled receptor protein signaling pathway DRD2, 1813G-protein coupled receptor protein signaling pathway ADRB2, 154G-protein coupled receptor protein signaling pathway CALM3, 808G-protein coupled receptor protein signaling pathway HTR2A, 3356G-protein coupled receptor protein signaling pathway DRD1, 1812G-protein signaling, coupled to cyclic nucleotide second messenger SSTR5, 6755G-protein signaling, coupled to cyclic nucleotide second messenger MTNR1A, 4543G-protein signaling, coupled to cyclic nucleotide second messenger CNR2, 1269G-protein signaling, coupled to cyclic nucleotide second messenger HTR6, 3362G-protein signaling, coupled to cyclic nucleotide second messenger GRIK2, 2898glutamate signaling pathway GRIN1, 2902glutamate signaling pathway GRIN2A, 2903glutamate signaling pathway GRIN2B, 2904glutamate signaling pathway ADAM10, 102integrin-mediated signaling pathway GRM7, 2917negative regulation of adenylate cyclase activity LRP1, 4035negative regulation of Wnt receptor signaling pathway ADAM10, 102Notch receptor processing ASCL1, 429Notch signaling pathway HTR2A, 3356serotonin receptor signaling pathway ADRB2, 154transmembrane receptor protein tyrosine kinase activation (dimerization) PTPRG, 5793transmembrane receptor protein tyrosine kinase signaling pathway EPHA4, 2043transmembrane receptor protein tyrosine kinase signaling pathway NRTN, 4902transmembrane receptor protein tyrosine kinase signaling pathway CTNND1, 1500Wnt receptor signaling pathway Many of the genes are indeed related to Alzheimer’s Disease through gamma secretase (presenilin) activity

9. Alan Ruttenberg School of Dental Medicine Nonsense Detection 9

10. Alan Ruttenberg School of Dental Medicine Say less 10

11. Alan Ruttenberg School of Dental Medicine Query 11

12. Alan Ruttenberg School of Dental Medicine Indexing 12

13. Alan Ruttenberg School of Dental Medicine ImmPort Semantic integration feasibility project ImmPort is an immunology database and analysis portal Goals: (meta)analysis, archiving, and exchange of scientific data Question: How can ontology and semantic technology help their mission Strategy: Choose a scientific problem of interest and explore how these technologies work in that context

14. Alan Ruttenberg School of Dental Medicine 3GSN Crystal structure of the public RA14 TCR in complex with the HCMV dominant NLV/HLA-A2 epitope

15. Alan Ruttenberg School of Dental Medicine 3gsn: zoom to emphasize functional features – contact points

16. Alan Ruttenberg School of Dental Medicine Challenge Curation of the sequence features, transfer via sequence alignment is tedious, error prone Information associating allele to disorder or disease is dispersed Difficult to do queries that link variation in alleles to disease and underlying biology Demonstrate how we can use semantic technologies to support access to knowledge for researchers on this project Opportunity

17. Alan Ruttenberg School of Dental Medicine Elements of representation PDB structures, translated into representation of molecular complex instance - “grain” of crystal – Main structure is via “part_of” relation ‘Canonical’ MHC molecule instances (should be classes) constructed from IMGT. Relate each residue in a PDB structure to the canonical residue, if one exists. Use existing ontologies (ChEBi, MaHCO, RO …) Annotate residues with various coordinate systems

18. Alan Ruttenberg School of Dental Medicine Elements of representation Contact points between peptide and other chains computed using JMOL w/5A diameter following IMGT. Represented as relation between residue instances. Structural features are (fiat) parts of molecules, and have residues as parts. Taken from SCOP, Pfam, SIFTS

19. Alan Ruttenberg School of Dental Medicine Visualization Prototype Highlighting all peptide contacts Highlighting residues that vary in this allele

20. Alan Ruttenberg School of Dental Medicine Variations in this allele select distinct ?hla_res where { ro:has_part ?hla_res; sc:blast_match ?allele. ?hla_res sc:reference_sequence_item ?item. ?allele sc:perturbation ?ares. ?ares sc:reference_sequence_item ?item. } SPARQL query results drive visualization Contact points in this HLA chain select distinct ?hla_res where { ro:has_part ?hla_res. ?hla_res sc:chain_contact ?peptide_res}

21. Alan Ruttenberg School of Dental Medicine Models of Development HCLS OBO Foundry W3C Working/Incubator/Interest Group Resource publication Aggregation 21

22. Alan Ruttenberg School of Dental Medicine How semantic web technologies help us achieve our goals (1) The project of enabling effective communication and discovery in the biological and life sciences is complex Some pieces of this effort are handled by these technologies We need tools for logical languages that have effective implementations. We get OWL, HerMIT, Pellet, FaCT++ In contrast to RDBs, tools work on combined schema (ontology) and data. o Both are expressed logically o Both are queryable together at once We get a free data model: RDF without any more effort beyond building the ontology o Saves us the trouble of designing a "data model" unless there is some good reason to do so. o Building "data models" is a common source of non-integrable data - everyone wants their own

23. Alan Ruttenberg School of Dental Medicine How semantic web technologies help us achieve our goals (2) We get databases that support those data models o Triple stores: Virtuoso, Stardog, OWLIM, more... Tools are standardized and there is a growing workforce we can tap. o This scales better than making our own tools, training people Our goal of universal easy access is shared with semweb efforts. The web is good at supporting this. o We need tools for logical languages that have effective implementations. We get OWL, HerMIT, Pellet, FaCT++ o By using these tools early we get to help guide their future

24. Alan Ruttenberg School of Dental Medicine How we help the semweb effort These standards and tools are still young Ontology building is hard. Too many semantic web efforts suffer from poor ontology design and the consequences of that - failure to integrated, confused messages. We help that by Providing and teach how to build high quality ontologies Provide a positive example of how a coordinated effort can develop good practices using the tools Provide feedback to standards efforts and tool developers that help improve successive versions Documenting our approaches so that others can benefit from them The semantic web needs us!