Carbbank data integration

[kmartinez834]

• [ ] Workflow to extract human records with UniProt accessions and PMIDs
• [ ] Mapping files

This task may become optional for 2.0 if we can not finish it

[kmartinez834]

From #115: This is about the integration of species, disease, tissue etc data into our system. This task may become optional for 2.0 if we can not finish it.

[kmartinez834]

Moved to 2.1 task list - will prioritize in May 2023

[kmartinez834]

@ubhuiyan

I created the following files (same content, different file types) to deal with the line breaks in the txt file. You can use these to help with mapping/instructions for Robel (the script I wrote is at /software/glygen/carbbank-mapping.py): generated/datasets/compiled/carbbank.csv generated/datasets/compiled/carbbank.json

I would aim to get the first 2 bullet points done this release, and try to work on tissue/cell line/disease mapping over time:

• "DB" has protein accessions (PIR, PIR1, PIR2, SwissProt) and publication ids (PMID)
• "BS" has organism, look for: "(CN) human","(CN) mouse","(CN) rat","(CN) pig","(CN) chicken","(GS) Saccharomyces cerevisiae","(GS) Dictyostelium discoideum","(GS) Drosophila melanogaster","(CN) fruit fly" --> Don't use "(CN) yeast", as this will result in many other types of yeast --> Tell Robel to use regex when parsing species and watch out for instances such as "(CN) human cytomegalovirus", "(CN) mouse-ear chickweed", or "(CN) human, mouse," (the later is a human-mouse xenograft)
• "BS" also has tissue, cell line and disease information. Need to populate the mapping file /data/projects/glygen/generated/misc/carbbank_mapping.csv , it includes all of the unique BS entries for the glygen organisms. --> In some cases, you have to look at multiple subfields, for example (disease) cancer, (OT) leukemia
• PMIDs for many records are available at GenomeNet (ex. https://www.genome.jp/entry/carbbank+4425) --> Need to discuss with Robel, how to use their SPARQL endpoint to get this information (RDF files for indiv records or bulk?)

[ReneRanzinger]

It would be good to know what the workflow is. In the end we want to assign species, tissue etc to glycans in the file.