Open kshefchek opened 6 years ago
As we do this, we should also refactor ingest to do make finer-grained distinctions between subtypes of of genetic interactions they capture. at present we map all genetic interaction relations to RO:0002435 (genetically interacts with), but BioGrid captures more detail here that would be useful (e.g. pos/neg genetic interactions, synthetic lethality interactions)
As we do this, we should also refactor ingest to do make finer-grained distinctions between subtypes of of genetic interactions they capture.
See http://version10.string-db.org/help/database/#table-networkactions
field | description |
---|---|
item_id_a | internal protein identifier. |
item_id_b | internal protein identifier. |
mode | type of interaction ("reaction", "expression", "activation", "ptmod"(post-translational modifications), "binding", "catalysis") |
action | the effect of the action ("inhibition", "activation") |
is_directional | no documentation |
a_is_acting | the directionality of the action if applicable (1 gives that item_id_a is acting upon item_id_b) |
score | the best combined score of all interactions in string. |
Thanks in part to @mellybelly @kltm @lwinfree and all the reusable data team, StringDB is now all cc-by, so we can pull a lot more data. Currently we are only ingesting the protein.links.detailed file.