DASH5 - nmdp-bioinformatics/dash GitHub Wiki

DaSH5: Berkeley, CHORI

In order to focus the activity at this event we have developed a list of “scrum-team” topics with tangible goals. We will form teams around some of these topics. The team should plan to produce code collaboratively and have a functional demo at the end of the event.

1. Prepare “feature service” for production use Implement a versioning system for the feature service Add authentication and security layers Add curation capability linking sequences to

• IPD-KIR and IMGT/HLA Database accession release version and Accession number (a more accurate way of saying the allele name)
• the sequence submitter and their “de-novo” identifiers for sequences they have submitted, associated with a given GFE notation/set of locus/feature/rank/accession coordinates Populate feature service with all versions of HLA.xml and KIR.dat Improve documentation; make sure that the documentation is sufficient to inform an API for analytics; expand on the BTOP-like pairwise difference annotation

2. GFE Service Enhancements

• Experiment with AWS configurations Add a database cache
• Volume test
• Test installation process
• Add clients and tools
• Rename (service-gfe-submission -> service-gfe) Connect to feature service with authentication

3. GFE Service Validation

• Validation of alignments using “features” from HLA.xml and KIR.dat files
• Develop automated tests; possibly using Neo4j

4. Improve current capabilities of the annotation pipeline (within GFE service)

• Make it a maven project; push code to maven central
• Allow GFE service to utilize different version of annotation
• Add capability of running ABO

5. Haplotype Frequency Curation Service (HFCu) development

• Build the shell of the backend based on the population service and check the code into GitHub

6. HL7 FHIR Develop use cases: • Vendor perspective (create partially filled FHIR template)

• identify what vendor software can pre-fill
• e.g. sequence info, allele-assignments, sample identification, methodology • Lab perspective
• fill in the rest of the data not available to vendor software • Transplant Center perspective
• patient demographics, from EMR • Registry perspective
• receive, process and verify bundle Identify required resources for use case Map data to resources • From HML
• build bundle
• POST to FHIR server Create client that will search, GET and link resources from FHIR server. Create a client that will access HLA terminology server.

7. Integrate Chipper algorithm for predicting proteasome cleavage sites

• Add to peptide prediction pipeline(s) https://github.com/massie/chipper

Educational Session

Since the HL7-FHIR topic involves a good deal of background we are proposing an educational session in a small breakout room (holds 7) with presentations by Bob Milius and Andrew Brown

• Intro to FHIR (Bob)
• ClinFHIR (general client to build FHIR resources and profiles) (Bob)
• Client to build FHIR resources from HML (Andrew)