Portable, workflow-based phenotype definitions

How do we achieve phenotype portability?
Download and immediately execute a phenotype against a local dataset.
 
Browse phenotypes represented in a standard format. 333 phenotypes are currently available.
 
Find and customise existing phenotypes, for local requirements.
 
Graphically author new phenotypes.

Phenotype execution only has two dependencies: cwltool and docker.

Why does phenotype portability matter?
Re-defining phenotypes at new sites may lead to inconsistent definitions. Instead, we want to be able to port definitions between sites, as facilitated by Phenoflow. This is not to say a single gold standard algorithm is required for each disease, but instead different algorithms should be semantically interoperable. The level of abstraction provided by Phenoflow also ensures this.
 
With the extraction of phenotype cohorts now forming a key part of many research methodologies, the mechanism used to derive a cohort within a given study needs to be reusable in order to ensure reproducibility and thus incremental research. The reusable definitions hosted within Phenoflow enable this research.
 
Trust in a phenotype definition is increased by validating it against multiple data sources, such as EHR repositories. By integrating with multiple data source formats, such as i2b2 and OMOP, Phenoflow enables this validation to take place.
 
Learn more
Phenoflow: A Microservice Architecture for Portable Workflow-based Phenotype Definitions (2021)
Martin Chapman, Luke V Rasmussen, Jennifer A Pacheco, Vasa Curcin