Interpretation of genomic data in Zebrafish: Relating Zebrafish Phenotypes to Human Disease Genes

Monte Westerfield, University of Oregon

Phenotype annotations of zebrafish (Danio rerio) genes that have known human disease orthologs provide a unique opportunity to analyze relationships between zebrafish and human phenotypes and to test the general applicability of these methods to other model organisms. Tools and methods being developed at the Center will be used to group human disease genes using Phenotype Ontology terms derived from OMIM descriptions. These same combinations of Phenotype Ontology terms can then be used to group zebrafish genes. We predict that each resulting group should contain a high proportion of orthologous human and zebrafish genes.

This system will also serve as a testing method for validating the algorithms, as failure to identify known orthologs could indicate that the algorithms are flawed. However, it could also indicate that the functions of the genes analyzed have largely not been conserved during evolution. We can distinguish between these possibilities by analysis of individual pairs of orthologs where sufficient biological information is available to determine independently whether gene function is conserved.

The methods we develop for carrying out these types of cross species phenotype comparisons can then be extended in many significant ways for future studies. For example, linking human and zebrafish (or other model organism) phenotypes in cases where the fish gene is not orthologous to the human gene can point toward candidates for human genes that are not yet known to be responsible for or to contribute to the human disease. Similarly, linking human and zebrafish phenotypes in cases where the molecular identity of the fish gene is not yet known, can provide candidates for the affected gene.

Although the Zebrafish Information Network (ZFIN) provides extensive information about zebrafish genetics and genomics, it currently provides only superficial descriptions of mutant and wildtype phenotypes. For this reason, we propose to develop robust support for phenotypes. We will curate published phenotype data and data submitted directly from large-scale mutant screens, subsequently using these data to annotate allele records in ZFIN and link them to human disease phenotypes.

We will also carry out a parallel annotation of the human homologs of the annotated zebrafish genes, using OMIM records as the informative source for this annotation. Finally, we will contribute ontologies and ontology extensions that we develop to OBO and we will deposit our annotations at the OBD to facilitate comparisons between these data and others.