Parsing of annotation / query by condition

[julien-roux]

(Issue originally opened at wirawara/BgeeDB#2)

It would be nice to have a function allowing to easily parse the Affymetrix and RNA-seq annotations data frames. For example it might be super useful to:

• query tissues and stages IDs using free-text strings
• for a given stage or organ ID, retrieve all parent/children stages/organs. This could be done with a webservice query very similar to what is done in loadTopAnatData.R to get the organ.relationships list.
• retrieve all samples annotated to a given organ / stage, including or not its sub-structures

[julien-roux]

See https://f1000research.com/articles/5-2748/v1#referee-response-18221, who also thinks this would be a good idea.

We could easily create:

• a getCallsByCondition function that would return all expression calls for the queried condition. This function would be very similar to loadTopAnatData.R, just outputing the gene2anatomy list in reverse order (anatomy2gene list)

• a getDataByCondition function that would return all processed data for chips / libraries (depending on the dataType of the input Bgee object) matching a queried condition: organ, stage, (optionally sex), (optionally strain). This function could easily use of a webservice query giving the list of all chip / libraries associated to this condition.
  What is a bit more tricky: to get access to the chips/libraries, the whole experiments .zip files would have to be downloaded, extracted, to use only one or a few files inside... Not very efficient. Another option would be to query the processed data trough the webservice: FTP file transfer success / processed data query through webservice #9

Something for the new postdoc to start with :p ? @marcrr @fbastian

[fbastian]

Should it also include sub-conditions? (e.g., expression in substructures of a queried organ). Or shold we provide relations between conditions in a different method?

[marcrr]

If it's easy from the R side, then yes I think that it's a good idea to include sub-conditions.

[julien-roux]

Ideally, the function could have an argument to specify this, and by default, yes I think it's a good idea to include sub-conditions