nlwashington
commented
10 years ago @cmungall , there are two ways i can generate reports: i can generate the report easily based on what is loaded into owlsim; basically do a computation over all individuals (or classes), and write out various summary stats. totally straightforward. however, this will miss the individuals that are not loaded directly. for example, any of the inferred diseases (intermediate nodes in the graph won't necessarily inherit their child phenotypes, and may not be loaded at all). also, all those genotypes or genes won't have. the way i've done it for the external analysis is to generate the ic scores separately, and then iterate over individuals without loading them into the graph itself. shall i continue with that method?
also, i think this will be much easier to complete for owlsim v3.
jmcmurry
i think it'd be nice to have some precomputed whole-system reports to show. these might be tables (downloadable) or could be pretty graphics, if desired. for example: survey of known disease models (mouse, fish, fly etc), and their phenotypic similarity scores, along with a list of their phenotypes in common, and other metadata like annot sufficiency and num of phenotypes. listing of annotated ontology classes (or not), direct or inferred overall listing of diseases, with metadata about num phenotypes, annot suff, etc. list of genes, and number of orthologs that have phenotypes
@cmungall should these just be some shell scripts that run against owlsim any time the server is started up (with new data)? they could spawn on jenkins and spit out static reports that are loaded into tables. they shouldn't be too hard to write.