The genetic cause underlying many human heritable diseases are still not identified. Model organisms with targeted modifications of the genome can help identifying the genetic causes for human diseases. Currently, there is a vast amount of gene-phenotype data across clinical resources, for model organisms with targeted modifications and non-model organisms. Combining and harmonizing these data, scattered across different databases, allows an integrated view of all the existing resources and helps identifying the genetic causes underlying a disease. Genetic diseases and the resulting animal model match are ranked by their phenotypic similarity to the disorder, thus constituting a prioritising resource.
Please see our Disease associations section to learn how these associations are performed.
The Phenogrid widget
The Phenogrid widget is displayed when you click on a plus sign associated to the disease, and it is developed by Monarch Initiaitve.
The Phenogrid widget allows to visualize the similarity results between:
- A set of phenotypes in a query (rows; in this case, phenotype annotations to a human disease)
- One or more phenotypically similar targets (in this case, phenotypes annotations to a mouse gene)
When a phenotype is shared between the query and the target, the intersection is coloured and the saturation of the cell correlates with the strength of the match. You can hover over the intersection to find the phenotype of the target and the query, and what is common between the two.
Phenodigm scores (Smedley et al. 2013) reflect the similarity between phenotype profiles, in this case the clinical descriptions of human diseases, as featured in OMIM, Orphanet and DECIPHER, described using the Human Phenotype Ontology (Köhler et al. 2017; HP terms), and mouse phenotypes mapped to the Mammalian Phenotype Ontology (MP terms). The raw score is normalised against the maximal possible score, which is the query matching to itself, so that the resulting score ranges between 0 and 100.
Associations are preformed in two ways
- Human diseases associated to a gene by orthology or direct annotation. Phenotype overlap between human disease-phenotype annotations and mouse gene-phenotype annotations, both harbouring mutations in the human and mouse orthologous gene.
- Human diseases predicted to be associated to Cib2 by phenotypic similarity. Phenotype overlap between human disease-phenotype annotations and mouse gene-phenotype annotations.
More on the Phenogrid widget and the matching algorithm
- Unmatched phenotypes. Phenotypes that were part of the query but are not shared by any of the targets can be seen by clicking the Unmatched Phenotype link.
- Two targets that share the same phenotypes may have a different overall score. This happens when there are phenotypes in the target that are not shared with the query. A target that matches the query for all phenotypes will rank higher than a target that matches all of them but also has one or more additional phenotypes that don’t match it; that is, the target is penalized for those phenotypes that are not sahred and thus ranks lower on the similarity scale.
Check out our Disease associations section
The Phenogrid widget can be downloaded from this github repository.
Visit the Monarch Initiative website