Operational

Members

• Mark Ebertowski - NMDP (mark.ebertowski@gmail.com)
• Steve Ozanne - NMDP (steve.ozanne@gmail.com)
• Matt Prestegaard - NMDP (GTalk - mpresteg@gmail.com)
• Scott Townsend - System Link - Histotrac
• Dan Valiga - NMDP
• Anthony Barber - NMDP
• Brad Baas - Illumina
• Kazu Osoegawa - Stanford Blood Center

Vision

To explore issues related to the use case of producing an HML message from a vendor passing through an integrator and arrive at a destination where the receiver can decode the message. Leverage what is learned through discussion to build services that do syntactic and semantic validation.

Problems Discussed

1. Data ends up at a computer where network connection not allowed.
2. Investigate the appetite for non relational databases.
3. GL String is not yet universal, operational needs are different than research needs.
4. Systems are coupled to NMDP allele codes.
5. Send in Genotype List and have a service respond if it valid at a biological level.
6. Only applicable for class II
7. Human consumption of HLA and allelic ambiguity
8. What is (or should be) in a test id?
9. Shouldn't there be the possibility to specify format for the consensus sequence?
10. How to specify test meta-data (e.g. Machine, Software Config on machine, assay brand/kit, test id, serial #)
11. Any attribute not in the HML specification could be implemented as a property. Is an extension (local schema) preferable to many new (and possibly inconsistent properties).

Definitions

Syntactic Validation

Validation that could be done by validating the HML document against the HML schema

Semantic Validation

Anything that can't be or isn't defined or restricted in the HML schema. (e.g. Data types, cross field validation)

Examples

Operational-Example-1-(NGS---Class-II)

URI of a consensus sequence

A standard would be needed for URIs of consensus sequence. Has this standard been developed?

1. How to resolve
2. Resolving
3. Where it came from
4. What the data is
5. HGS
6. Quality
7. Bases
8. Region