SHIRAZ and CABERNET: Leveraging automation, crowdsourcing, and ontologies to improve the accuracy and throughput of zebrafish histological phenotype annotations

Brian Canada, Georgia Thomas, John Schleicher, James Z. Wang, Keith C. Cheng

Research output: Contribution to journalConference article

Abstract

One of the goals of the Zebrafish Phenome Project is to systematically annotate the cellular-level morphological phenotypes associated with each gene in the zebrafish genome. Here, we offer demonstrations of two complementary software tools designed to help achieve that objective: SHIRAZ, a content-based image retrieval system designed for automated high-throughput annotation of histological phenotypes in the larval zebrafish, and CABERNET, a "crowdsourcing" application for histology image tagging that enables multiple domain experts to achieve consensus on ontology-compliant phenotype annotations. Potentially, such "consensus annotations" not only can be used to improve the accuracy of ground truth data for training SHIRAZ. but they can also be imported directly into PATO- compatible phenomic databases such as the Zebrafish Information Network.

Original languageEnglish (US)
Pages (from-to)288-289
Number of pages2
JournalCEUR Workshop Proceedings
Volume833
StatePublished - Dec 1 2011
Event2nd International Conference on Biomedical Ontology, ICBO 2011 - Buffalo, NY, United States
Duration: Jul 26 2011Jul 30 2011

All Science Journal Classification (ASJC) codes

  • Computer Science(all)

Fingerprint Dive into the research topics of 'SHIRAZ and CABERNET: Leveraging automation, crowdsourcing, and ontologies to improve the accuracy and throughput of zebrafish histological phenotype annotations'. Together they form a unique fingerprint.

  • Cite this