Using the phenoscape knowledgebase to relate genetic perturbations to phenotypic evolution

Prashanti Manda, James P. Balhoff, Hilmar Lapp, Paula Mabee, Todd Vision

Abstract

The abundance of phenotypic diversity among species can enrich our knowledge of development and genetics beyond the limits of variation that can be observed in model organisms. The Phenoscape Knowledgebase (KB) is designed to enable exploration and discovery of phenotypic variation among species. Because phenotypes in the KB are annotated using standard ontologies, evolutionary phenotypes can be compared with phenotypes from genetic perturbations in model organisms. To illustrate the power of this approach, we review the use of the KB to find taxa showing evolutionary variation similar to that of a query gene. Matches are made between the full set of phenotypes described for a gene and an evolutionary profile, the latter of which is defined as the set of phenotypes that are variable among the daughters of any node on the taxonomic tree. Phenoscape’s semantic similarity interface allows the user to assess the statistical significance of each match and flags matches that may only result from differences in annotation coverage between genetic and evolutionary studies. Tools such as this will help meet the challenge of relating the growing volume of genetic knowledge in model organisms to the diversity of phenotypes in nature.

Introduction: Bridging the Gap Between Genetics and Natural History

Biological data is currently siloed into two distinct domains. Developmental geneticists study "model organisms," such as mice and zebrafish, to observe how specific genetic perturbations—like gene knockouts—alter an animal's form. Conversely, evolutionary biologists and paleontologists have spent centuries documenting the diverse morphologies of thousands of species across the tree of life.

A significant barrier to integrating these fields is linguistic. A geneticist might describe a mutant fish as having "reduced pelvic fins," while a taxonomist might describe a related species as "lacking a pelvic girdle." Because these descriptions are traditionally recorded in free-form natural language, searching across these datasets to identify the genetic basis of evolutionary shifts has been nearly impossible. The Phenoscape Knowledgebase (KB) addresses this by translating natural language descriptions into a machine-readable format, enabling automated "similarity searches" that link genetic mutations to evolutionary changes.

The Phenoscape Framework

The Phenoscape KB relies on three core pillars to create a unified data environment:

  1. Standardized Ontologies: To establish a common language, the KB utilizes ontologies such as Uberon (for anatomy), PATO (Phenotype and Trait Ontology for qualities like shape or size), and the Vertebrate Taxonomy Ontology.
  2. Entity-Quality (EQ) Annotations: Natural language descriptions are converted into structured data. For example, a description of "large teeth" is decomposed into an Entity (tooth) paired with a Quality (increased size).
  3. Evolutionary Profiles: Rather than focusing solely on individual species, the KB uses a parsimony algorithm (Fitch’s algorithm) to infer where a trait changed on the evolutionary tree. An "evolutionary profile" represents the set of traits that vary between the descendants of a specific taxonomic node.

Computational Discovery via Semantic Similarity

To compare a gene's effects to evolutionary changes, the system calculates a semantic similarity score. This process utilizes Information Content (IC) based on the Most Informative Common Ancestor (MICA) within the ontology. This approach functions as a "BLAST for phenotypes," identifying biologically relevant matches even when descriptions are not identical, provided they refer to related structures or qualities (e.g., recognizing that a "premaxilla" and a "maxilla" are both components of the upper jaw).

To ensure statistical rigor, the KB implements Expect (E) values, a concept borrowed from bioinformatics. This metric indicates the likelihood that a match between a gene and a species occurred by chance, helping researchers filter out noise and identify high-confidence candidate genes.

Data Integration and Scale

As of the current implementation, the Phenoscape KB integrates a massive volume of heterogeneous data:

  • Model Organism Data: Phenotype profiles for genes sourced from Mouse Genome Informatics (MGI), Xenbase (frogs), ZFIN (zebrafish), and the Human Phenotype Ontology.
  • Evolutionary Data: Annotations for over 4,800 species and higher taxa derived from 149 published phylogenetic studies, encompassing over 19,000 distinct character states.
  • Precomputed Profiles: The KB contains 636 precomputed evolutionary phenotype profiles, targeting the specific "moments" in evolution where morphology shifted.

Case Study: Wnt7a and Limb Evolution

The utility of the KB is demonstrated through a case study of the mouse gene Wnt7a, which is known to be involved in limb development. When Wnt7a was queried against the KB, the suborder Microchiroptera (bats) emerged as the top evolutionary match.

The match yielded an E-value of 0.06, suggesting a strong, non-random similarity. While there were no "exact" matches in the traditional sense, the system identified a high number of similar phenotypic pairs across both upper and lower limbs. This supports the hypothesis that Wnt7a is a strong candidate gene for studying major morphological transitions, such as the fin-to-limb transition or the specialized development of bat wings.

Limitations and Future Directions

While the Phenoscape KB represents a major step forward, several challenges remain:

  • Annotation Bias: The database is currently weighted toward skeletal anatomy and specific groups like fishes. Traits involving soft tissue or behavior are underrepresented due to the difficulty of curating them from existing literature.
  • Curation Bottlenecks: Manual curation remains a significant bottleneck, and the KB is currently an incomplete sample of known vertebrate variation.
  • Biological Complexity: A high similarity score suggests that the same developmental pathways may be involved, but it does not definitively prove a gene caused an evolutionary change. Factors like convergent evolution or different genes within the same pathway must be considered.

Future work will focus on incorporating more complex semantic similarity measures and developing automated tools to accelerate the curation of natural language into structured ontologies. The Phenoscape KB is available for public use at http://kb.phenoscape.org.

Summary updated

Abstract

The abundance of phenotypic diversity among species can enrich our knowledge of development and genetics beyond the limits of variation that can be observed in model organisms. The Phenoscape Knowledgebase (KB) is designed to enable exploration and discovery of phenotypic variation among species. Because phenotypes in the KB are annotated using standard ontologies, evolutionary phenotypes can be compared with phenotypes from genetic perturbations in model organisms. To illustrate the power of this approach, we review the use of the KB to find taxa showing evolutionary variation similar to that of a query gene. Matches are made between the full set of phenotypes described for a gene and an evolutionary profile, the latter of which is defined as the set of phenotypes that are variable among the daughters of any node on the taxonomic tree. Phenoscape's semantic similarity interface allows the user to assess the statistical significance of each match and flags matches that may only result from differences in annotation coverage between genetic and evolutionary studies. Tools such as this will help meet the challenge of relating the growing volume of genetic knowledge in model organisms to the diversity of phenotypes in nature. The Phenoscape KB is available at http://kb.phenoscape.org.
BibTeX
@inproceedings{manda2015using,
  author = {Prashanti Manda and James P. Balhoff and Hilmar Lapp and Paula Mabee and Todd Vision},
  title = {{Using the phenoscape knowledgebase to relate genetic perturbations to phenotypic evolution}},
  booktitle = {Genesis},
  year = {2015},
  doi = {10.1002/dvg.22878},
}
Figure