Integrating data with phylogenies, at scale
0 likes455 views
The document discusses the challenges of integrating nomenclature and phylogenetic data, highlighting the complexity of reconciling names and concepts within biological taxonomy. It emphasizes the importance of phyloreferencing as a systematic approach to mapping phylogenetic knowledge, providing a reliable navigation method for data integration. The authors argue for the development of scalable ontologies and reasoning techniques to enhance our understanding of the tree of life.
Integrating data with phylogenies, at scale
- 3. What’s in a name? Chaos! • Names and Concepts do not reconcile that easily • Names are text strings • Context is lacking or subjec(ve • Meaning is not computable
- 8. Idiosyncratic Russian dolls syndrome
- 9. Idiosyncratic Russian dolls syndrome
- 10. Idiosyncratic Russian dolls syndrome
- 11. Idiosyncratic Russian dolls syndrome
- 12. Idiosyncratic Russian dolls syndrome
- 13. Idiosyncratic Russian dolls syndrome
- 14. Idiosyncratic Russian dolls syndrome
- 19. Dark taxa!
- 23. Tree-thinking Berberidopsidaceae Opiliones Zingiberaceae Hamamelidaceae Sarcolaenaceae Lingulidae Hymenoptera Mammalia Apocynaceae Galliformes Rubiaceae Anarthriaceae Lineidae Crocodylidae Stylosiphonia Andrenidae Cracidae Gavialis Globba Micrella Rhodoleia Phalangiidae Tachyglossa Lyginia Mediusella Chamaeclitandra
- 24. Tree-thinking Berberidopsidaceae Opiliones Zingiberaceae Hamamelidaceae Sarcolaenaceae Lingulidae Hymenoptera Mammalia Apocynaceae Galliformes Rubiaceae Anarthriaceae Lineidae Crocodylidae Stylosiphonia Andrenidae Cracidae Gavialis Globba Micrella Rhodoleia Phalangiidae Tachyglossa Lyginia Mediusella Chamaeclitandra These names are not generated in an evoluDonary-based framework (Groups defined by character similarity vs. common descent)
- 32. FEED Textual Definition – The hyoglossus is a muscle that attaches to the hyoid and tongue and is innervated by Cranial Nerve XII. Computable Definition – ('attached to' some 'hyoid bone') and ('attached to' some tongue) and ('innervated by' some 'hypoglossal nerve') and spatially disjoint with 'intrinsic tongue muscle' Druzinsky et al (2015): Logic definiDons of mammalian feeding muscles by means of necessary and sufficient condiDons true for all mammals Nomenclature ≠ Seman(cs
- 34. Phyloreferences Statements formally expressing the paaerns we discover (analogous to map coordinates) Node-Based Branch-Based Apomorphy-Based A B C A B C A B C X The clade originaDng with the last common ancestor of B and C. The clade originaDng with the first ancestor of B that is not an ancestor of A. The clade originaDng with the first ancestor of C to evolve X.
- 35. Phyloreferences yield a coordinate system for the Tree of Life • Any node, branch, subtree is referenceable • References are unambiguous • References are computable • References are portable • Adapts to new and changing knowledge
- 36. Many needed technologies already exist • OWL ontologies designed for – PhylogeneDc knowledge: CDAO – Phenotypic knowledge: Uberon, PATO, … – Efficient and expressive reasoners: FaCT++, HermiT, Racer, ELK
- 40. Phyloreferences as ontological expressions Phyloreference expressions can be: • Easily generated by anyone • Can work on any tree • Named and registered – To promote reuse and consistency – To improve usability and accessibility Class: Campanulaceae Annota(ons: rdfs:label “Campanulaceae_aier_1995” dc:descripDon “the clade that includes Campanula laDfolia but not Sphenoclea” EquivalentTo: cdao:has_Descendant value taxon:Campanula_laDfolia and phyloref:excludes_lineage value taxon:Sphenoclea Class: AGF4-SHRU-3560 EquivalentTo: cdao:has_Descendant value taxon:Campanula_laDfolia and phyloref:excludes_lineage value taxon:Sphenoclea vs.
- 41. Challenges • OWL-based data model to saDsfy phylogeneDc taxonomy, reasoning expressivity, scalability • ConvenDons for data transformaDon, and consequences of different choices • Least common ancestor reasoning for OWL data • Lack of canonical specimen idenDfier system • Specifier mapping ontologies
- 42. Tree of Life, ontologized: A universal coordinate system • The Tree of Life is itself an aggregaDon and integraDon of our phylogeneDc knowledge. • Phyloreferencing is addressing into a knowledge universe. • Ontologies, reasoning, and other KR techniques are powerful tools for this.
- 43. Acknowledgements • NaDonal Science FoundaDon (DBI-1458484) • Ken and Linda McGurn • Phenoscape • EvoIO