clemente-lab / mmeds-meta

A database for storing and analyzing omics data
https://mmeds.org
2 stars 1 forks source link

Remove old sub-analysis functionality and re-integrate as a variant of meta-analysis #431

Open adamcantor22 opened 1 year ago

adamcantor22 commented 1 year ago

Is your feature request related to a problem? Please describe. There is a lot of code in MMEDS dedicated to the handling of "sub-analysis" or "child analysis". This is generally handled inside config files, but not much is present besides the framework. Past this, we have been upgrading MMEDS without regard for these functionalities, and they are likely broken or incompatible with our current methods of analysis. However, we still need an easy way to do "sub-analysis", where we run an analysis tool on a specific subset of one study.

Describe the solution you'd like A sub-analysis can simply be a type of meta-analysis that is contained to one study, i.e. if a meta-study query would have been WHERE SpecimenBodySite='Gut', the query for a sub analysis of the same kind for a given study ExampleStudy would be WHERE SpecimenBodySite='Gut' AND StudyName='ExampleStudy'. This will require only minor modifications to the current meta-analysis pipeline, most significantly allowing the sub-analysis to exist as a directory inside the original study directory. One specific change could be this: there is a boolean variable now meta_study that is true when a meta-study is being uploaded. Replace this with a variable study_type that can have values 'original', 'meta', or 'sub'.

Additional Context The old child analysis code should be removed, which will be a significant improvement to codebase readability.