Computing on Phenotypes AMP 2015
Download as PPTX, PDF1 like3,837 views
1. The document discusses using phenotypes across species to aid in interpreting genomic data from patients and improving diagnosis and treatment. 2. Building comprehensive phenotype databases from multiple sources is challenging due to disparate data on human genes/variants and model organisms. 3. The Monarch Initiative aims to link human diseases to phenotypes in model systems through an ontology-based knowledge base and portal. 4. Incorporating rich phenotypic data can improve variant filtering and interpretation by providing more context for sequencing results.
Computing on Phenotypes AMP 2015
- 1. Computing on phenotypes across scale and species Chris Mungall, PhD AMP 2015 @monarchinit @chrismungall
- 3. Patient Genome /Exome Improved Diagnosis, treatment filtering * ** ***** **** Phenome Gene to Phenotype Database Genomic data Hyperkeratosis, hearing impairment, …
- 4. Obstacles to phenome-based interpretation • Building a comprehensive phenomic database – Multiple disparate sources: • Human Genes, Variants, etc databases • Orthologous genes in model organisms • Phenotype Search and Matching • How do utilize phenotypes in a variant filtering pipeline? • How do we match phenotypes in different species? • How much difference does phenotype make?
- 5. monarchinitiative.org Interpretation requires prior knowledge of gene-phenotype effects
- 6. monarchinitiative.org Model organisms supply ~50% phenotypic knowledge to human genes
- 7. Other organisms provide deeper molecular pathological perspective SNCA (Hsap) phenotypes • Mental deterioration • Urinary urgency • Lewy bodies • Tremor • Urinary Urgency • Substantia nigra gliosis • … Snca (Mmus) phenotypes • Retinal dopaminergic neuron degeration (OMIM) (MGI) • Abnormal synaptic dopamine release • Alpha-synuclein inclusion body • Dopamine neuron loss • … Transgenic Snca (Dmel) phenotypes (FlyBase)
- 8. Monarch Portal: linking human diseases to model systems • One stop shop for gene-phenotype data and analysis: • Humans • Models – Data: • Genes • Variants • Complex genotypes • Phenotypes • Disease http://monarchinitiative.org/ Mungall, C. J., Washington, N. L., Nguyen-Xuan, J., Condit, C., Smedley, D., Köhler, S., … Haendel, M. A. (2015). Use of Model Organism and Disease Databases to Support Matchmaking for Human Disease Gene Discovery. Human Mutation, 36(10), 979– 84. doi:10.1002/humu.22857
- 9. monarchinitiative.org Building the knowledge base + in-house curation Phenotypes From 60 metazoan species
- 10. How do we search phenome databases? • Given a patient phenotypic profile • What are the relevant genes implicated in… – Humans? – Model systems? Patient Phenome Gene <-> Phenotype Database Hyperkeratosis, hearing impairment, … Candidate genes KRT2 GJB2
- 11. monarchinitiative.org We have a common computable language for sequence data…. ATCTTAGCACGTTAC… OR g.241T>c ….not so much for phenotypes
- 13. monarchinitiative.org Ulcerated paws Palmoplantar hyperkeratosis Thick hand skin Abnormal autopod skin Ontologies: Concepts Inter-related in a graph Hyperkeratosis
- 14. monarchinitiative.org Ulcerated paws Palmoplantar hyperkeratosis Thick hand skin Abnormal autopod skin id: HP:0000972 Synonyms: “Thick palms and soles” Def: “Hyperkeratosis affecting the palm of the hand and the sole of the foot” Köhler, S., Doelken, S. C., Mungall, … Robinson, P. N. (2013). The Human Phenotype Ontology project: linking molecular biology and disease through phenotype data. Nucleic Acids Res., Kohler, S.(1), gkt1026–. doi:10.1093/nar/gkt1026 OMIM:309560 OMIM:613989 … MP:0000578 Ctsk Ntrk1 Lamc2
- 15. monarchinitiative.org Ulcerated paws Palmoplantar hyperkeratosis Thick hand skin Abnormal autopod skin id: HP:0000972 Synonyms: “Thick palms and soles” Def: “Hyperkeratosis affecting the palm of the hand and the sole of the foot” OMIM:309560 OMIM:613989 … ? Ctsk Ntrk1 Lamc2 MP:0000578
- 16. monarchinitiative.org paw skin hand autopod = epidermis stratum corneum Mungall, C. J., Torniai, C., Gkoutos, G. V, Lewis, S. E., & Haendel, M. A. (2012). Uberon, an integrative multi-species anatomy ontology. Genome Biology, 13(1), R5. doi:10.1186/gb-2012-13-1-r5 Keratinization (GO) Uberon bridges multiple ontologies
- 17. monarchinitiative.org Ulcerated paws Palmoplantar hyperkeratosis Thick hand skin Abnormal autopod skin id: HP:0000972 Synonyms: “Thick palms and soles” Def: “Hyperkeratosis affecting the palm of the hand and the sole of the foot” Ctsk Ntrk1 Lamc2 MRCA Phenotype What model organism genes are relevant for my phenotype?
- 18. monarchinitiative.org Smedley, D., Oellrich, A., Köhler, S., Ruef, B., Westerfield, M., Robinson, P., … Mungall, C. (2013). PhenoDigm: analyzing curated annotations to associate animal models with human diseases. Database : The Journal of Biological Databases and Curation, 2013, bat025. doi:10.1093/database/bat025 Multi-phenotype search
- 19. Smedley, D., Oellrich, A., Köhler, S., Ruef, B., Westerfield, M., Robinson, P., … Mungall, C. (2013). PhenoDigm: analyzing curated annotations to associate animal models with human diseases. Database : The Journal of Biological Databases and Curation, 2013, bat025. doi:10.1093/database/bat025
- 20. monarchinitiative.org PhenoGrid phenotype comparison widget Patient phenotypes Compare patients with: Other patients Known diseases Models http://monarchinitiative. org/page/phenogrid
- 21. PHenotypic Interpretation of Variants in Exomes Whole exome Remove off-target and common variants Variant score from allele freq and pathogenicity Phenotype score from phenotypic similarity (hi)PHIVE score to give final candidates Mendelian filters https://www.sanger.ac.uk/reso urces/software/exomiser/
- 22. monarchinitiative.org Adding phenotype improves variant interpretation Robinson, P., Kohler, S., Oellrich, A., Wang, K., Mungall, C., Lewis, S. E., … Köhler, S. (2013). Improved exome prioritization of disease genes through cross species phenotype comparison. Genome Research. doi:10.1101/gr.160325.113
- 23. monarchinitiative.org Patient diagnosis example Deleteriousness Phenotype Score P ID Gen e MT P2 S Clinical Pheno Matching Pheno gene P Var ES Ran k 92 9 SMS 1.00 0.99 0.00 Ostopenia Decreased BMD Sms 0.4 1.00 0.89 1/25 Short stature Decreased body length Neonatal hyoglycemia Decreased circulating glucose levels acidosis Decreased circulating potassion levels Decreased body weight Decreased body weight Bone, W. P. et al. Computational evaluation of exome sequence data using human and model organism phenotypes improves diagnostic efficiency. Genet. Med. in press, (2015)
- 24. monarchinitiative.org From exomes to genomes Smedley D. et al, under review
- 25. Building up a massive phenomic database • Initial efforts • Manual curation of OMIM records • Expert biocurators and clinicians • Lag between publication and phenotype capture • How are we scaling up? • Phenotypes at time of publication • Working with patient registries • Natural Language Processing • Integration with Gene Ontology curation
- 26. Each case Report Associated With HPO profile Robinson, P. N., Mungall, C. J., & Haendel, M. (2015). Capturing phenotypes for precision medicine. Molecular Case Studies, 1(1), a000372. doi:10.1101/mcs.a000372
- 27. Beyond mendelian phenotypes • First pass • Mendelian or ‘rare’ diseases • Can we include a broader definition of ‘phenotype’ • Quantitative traits, e.g. hippocampus volume • Common disease phenotypes • Cancer
- 28. monarchinitiative.org Groza, T., … Robinson, P. N. (2015). The Human Phenotype Ontology: Semantic Unification of Common and Rare Disease. The American Journal of Human Genetics, 1–14. doi:10.1016/j.ajhg.2015.05.020 Mining pubmed for phenotypes F-Score: 45%
- 30. Building causal molecular pathological models http://create.monarchinitiative.org http://noctua.berkeleybop.org
- 31. Conclusions • Phenotypes are crucial for precision medicine • Variant interpretation needs more than genome data • Methods of incorporating phenotypes are evolving • We need all the organisms • The Monarch Portal integrates and organizes gene-phenotype data • Ontologies make phenotypes computable • Depth and breadth of structured phenotype data is growing
- 32. Monarch team Lawrence Berkeley Chris Mungall Nicole Washington Suzanna Lewis Jeremy Nguyen Seth Carbon Charité Peter Robinson Sebastian Kohler Max Schubach Tomasz Zemojtel U of Pittsburgh Harry Hochheiser Mike Davis Joe Zhou OHSU Melissa Haendel Nicole Vasilesky Matt Brush Kent Shefchek Julie McMurry Mark Engelstead Sanger Institute Damian Smedley Jules Jacobson Garvan Tudor Groza Craig McNamara Edwin Zhang Funding: NIH Office of Director: 1R24OD011883 NIH-UDP: HHSN268201300036C, HHSN268201400093P http://monarchinitiative.org
- 34. From phenomes to exposomes • Environmental context • Microbiome • Drugs Buttigieg, P. L., Morrison, N., Smith, B., Mungall, C. J., & Lewis, S. E. (2013). The environment ontology: contextualising biological and biomedical entities. Journal of Biomedical Semantics, 4(1), 43. doi:10.1186/2041-1480-4-43
Editor's Notes
- #2: Phenotype-Gene Associations in Variant Interpretation
- #3: Even with increased genomic data e.g. EXAC, it can still be hard to pinpoint the causative variant, or to be sure the real variant hasn’t been filtered
- #6: Human: GWAS, OMIM, clinvar Orthology via PANTHER v9 When put together, they bring the phenotypic coverage of human genes (either directly or inferred via orthology) up to nearly 80%. That is A LOT of coverage. How can we better tap that?
- #7: Human: GWAS, OMIM, clinvar Orthology via PANTHER v9 When put together, they bring the phenotypic coverage of human genes (either directly or inferred via orthology) up to nearly 80%. That is A LOT of coverage. How can we better tap that?
- #8: More molecular. Just as the exome is only an incomplete view of the genome, only part of the phenome is observed and measured
- #12: We can quantify distance due to shared ancestry. Draw trees.
- #15: Mention HPOA
- #16: Mention HPOA
- #17: Backbone ontology. Bridges anatomical and pathological levels
- #33: There are a lot of people who have contributed to this work over many years.
- #38: If we include bridging ontologies, we can unify diseases across sources AND phenotypes across sources and organisms.
- #39: Represent Human as a biological subject Represent diseases as collections of nodes in the graph 3. Interoperable with other bioinformatics resources and leverage modern semantic standards