I was able to simplify the ingest a little by using slightly different files. We're using their gene file to make gene nodes now, and they have a g2(m2)p file that includes the gene ids, which eliminated one mapping step. I was also able to remove the NCBI gene bits, since we just want to stick with dictyBase's own gene IDs.
I also managed to squeeze the phenotype name to id step into our infrastructure more easily than the robot approach by using the semantic sql version of the dictybase phenotype ontology. (thank you @cmungall for the tutorial session & great docs!)
I was able to simplify the ingest a little by using slightly different files. We're using their gene file to make gene nodes now, and they have a g2(m2)p file that includes the gene ids, which eliminated one mapping step. I was also able to remove the NCBI gene bits, since we just want to stick with dictyBase's own gene IDs.
I also managed to squeeze the phenotype name to id step into our infrastructure more easily than the robot approach by using the semantic sql version of the dictybase phenotype ontology. (thank you @cmungall for the tutorial session & great docs!)