Only parse biological_process

[JasonTan-code]

Hi,

Thanks for the code. It seems this obo_parser can only work on biological_process and ignore cellular_component and molecular_function.

[allenbaron]

obo_parser is designed to extract records descending from a single root node. If not specified with the root_id argument obo_parser will identify a root node using the records and then extract all of its descendants. The three gene ontology terms you specify do not share a parent term; each serves as a root node. That's why it only ever downloads biological_process. You can deal with this in one of 3 ways:

1. Add a record to the .obo file that represents an arbitrary parent node for the 3 terms and add that as the parent for each of the GO root nodes (biological_process, cellular_component, and molecular_function).
2. Run obo_parser 3 times specifying each of biological_process (GO:0008150), cellular_component (GO:0003674), and molecular_function (GO:0005575) as the root_id once. This will generate 3 files, dropping "obsolete" entries, which you could then combine.
3. Use a fork of obo_parser that I modified to extract all the records; it is pretty much the same but accepts a --return_all flag. This will keep all records, including "obsolete" entries. Note: I haven't made any updates to correct url errors so you'll have to download the file first.