transcriptomic query functions #105
Conversation
There was a problem hiding this comment.
-
Main method looks good - just small query is returning false rather than empty dataframe when no results are returned consistent with behaviour of other methods on VFB_connect?
-
Not sure why changes were made to lookup. I think these are likely to break other methods.
Please revert (or explain the motivation so we can discuss other fixes). -
Having played with running queries I am concerned that the results can appear confusing when multiple clusters are annotated with the same ontology term in a single paper. We need to look at returning more context in these cases. Example:
| cell_type | cell_type_id | ref | gene | gene_id | function | extent | level |
|---|---|---|---|---|---|---|---|
| adult Kenyon cell | FBbt_00049825 | Li et al., 2022, Science 375(6584): eabk2432 | 14-3-3ε | FBgn0020238 | [] | 0.678571 | 1268.15800 |
| adult Kenyon cell | FBbt_00049825 | Li et al., 2022, Science 375(6584): eabk2432 | 14-3-3ε | FBgn0020238 | [] | 0.705882 | 1257.36780 |
| adult Kenyon cell | FBbt_00049825 | Mokashi et al., 2021, Front. Psychiatry 12: 69... | 14-3-3ε | FBgn0020238 | [] | 0.805285 | 1712.31100 |
| adult Kenyon cell | FBbt_00049825 | Mokashi et al., 2021, Front. Psychiatry 12: 69... | 14-3-3ε | FBgn0020238 | [] | 0.823817 | 1532.72750 |
| adult Kenyon cell | FBbt_00049825 | Mokashi et al., 2021, Front. Psychiatry 12: 69... | 14-3-3ε | FBgn0020238 | [] | 0.816967 | 1598.16420 |
We could just return the cluster name as an additional column, but it would be better if we could return more informative context. In these cases the difference appears to be sex.
src/vfb_connect/neo/neo4j_tools.py
Outdated
| @@ -237,7 +237,7 @@ def get_lookup(self, limit_by_prefix=None, include_individuals=False, | |||
| q = self.commit_list([property_lookup_query]) | |||
| out.extend(dict_cursor(q)) | |||
| lookup = {x['name']: x['id'].replace('_', ':') for x in out} | |||
| lookup.update({x['id']: x['id'].replace('_', ':') for x in out}) | |||
There was a problem hiding this comment.
Why change this? I think the original design makes sure curie inputs are lookups for short_forms. Pretty sure other methods depend on this. The changed code looks like it updates a library with they keys and values it was initialised with.
There was a problem hiding this comment.
This fixed an issue I was having - but now unable to replicate! changed back.
This reverts commit d0ebf52.
Yeah, happy to change - I think methods on VFB_neo4j usually return False rather than empty datastructures? I didn't want to throw an error, because having no data isn't really an error, but empty dataframe makes sense too. |
I think either option is fine. We should just be consistent across methods. |
|
I think neo_query_wrapper doesn't return the headers if there is no data, so would need to manually specify if we want to return a dataframe with column headers and no data |
|
copied return from get_connected_neurons_by_type |
|
Looks good. Happy to merge. BTW We should think about how to deal with mixed sex => blank. |
I don't think there is any annotation in FlyBase to distinguish mixed sex from unknown sex |
fixes #104