- Created by Bret Heale, last modified on May 03, 2022
| Feedback | During Connectathon, users can capture feedback related to the operations and/or the reference implementation here. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Short Description | Using/accessing FHIR Genetic data - this track is about what/how you can access data to meet clinical use case. Operations expand the capabilities of a FHIR server to deliver genomics functionality to answer key clinical queries in a streamlined fashion, that is agnostic to the format used to report genetic data (i.e. the input parameters use a single format modality rather than an application needing to know a priori how the data is stored in the server). New operations are proposed that return associated Diagnostic or Therapeutic implications – answering such a request as ‘return the diagnostic implications associated with a specific genetic variation’ or 'return the diagnostic and therapeutic implications of patients that match my patient genetically' without needing to know the variant formatting on the server. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Long Description | Operations expand the capabilities of a FHIR server to deliver genomics functionality. Even with conformance to the Clinical Genomics IG, genetics can require dealing with liftover, synonymous HGVS, subsumed variants, overlapping variants, mapping between system representations, and multiple formatting modalities. Application developers, or other users of genetic data, will appreciate how operations simplify development effort, providing a single input format and streamline access. This track is about using FHIR genomics data. Through understanding the gaps in usability we will build a better more usable standard. Without an operation, FHIR queries for genetic data must take into account multiple synonymous HGVS, liftover, subsumed variants and multiple formatting modalities (see https://build.fhir.org/ig/HL7/genomics-reporting/branches/operations/operations.html#general-guidance-for-the-use-of-these-operations). Key clinical queries such as ‘find all patients with genetic variants like my patient,’ ‘does my patient have a specific variant?’, ‘does my patient have a variation in a specific gene of interest?’, and ‘does the patient’s tumor qualify for a specific trial’ are all examples of queries that are simplified using operations. New operations are operations proposed that can answer the question ‘do any of my patient’s genetic variations have diagnostic or therapeutic implications?’ We invite you to try these operations out using a reference implementation of the operations with data from oncology, pharmacogenomic and general genomic knowledge resources. Use cases such as pharmacogenomics reanalysis, therapeutic treatment recommendations, ACMG screening for common conditions and population – cohort – based retrieval are now supported. Our goal is to try out as many of the proposed operations as possible. We are seeking feedback, including utility of responses, additional parameters, speed of development (with/out operations), coverage of use cases, gaps, additional operations needed. While scenarios are presented here, we are eager to work with/support you on your use cases. We invite application developers, EHR vendors, Business and Data analysts/scientists, especially, to make use of the reference server. See the scenario section for example scenarios, additional scenarios are welcome during the connectathon. If one would like to develop their own server that is in-scope and we will be happy to try out your server. Additionally, if a knowledge-base, e.g. PharmGKB, would like to test out delivering their content using the Implication operations that would be encouraged. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Type | Test a FHIR-associated specification | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Submitting Work Group/Project/Accelerator/Affiliate/Implementer Group | This track will be hosted by CodeX's GenomeX FHIR Accelerator project (Operations Use Case) in collaboration with the Clinical Genomics Work Group with input from any interested parties. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Track Lead(s) | Bob Dolin BDolin@elimu.io Bret Heale bheale@humanizedhealthconsulting.com | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Track Lead Email(s) | |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Tracks |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
FHIR Version | Current Build, FHIR R4 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Specification(s) this track uses | Currently At: https://build.fhir.org/ig/HL7/genomics-reporting/branches/operations/operations.html Will be at https://build.fhir.org/ig/HL7/genomics-reporting/operations.html
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Artifacts of focus | https://build.fhir.org/ig/HL7/genomics-reporting/branches/operations/operations.html Reference Implementation: https://fhir-gen-ops.herokuapp.com/ Reference implementation (open-source) git-repo: https://github.com/FHIR/genomics-operations Postman library: https://github.com/FHIR/genomics-operations/blob/main/FHIRGenomicsOperations.postman_collection.json Description of Reference Implementation (patient lists, capabilities, etc... of reference server): https://docs.google.com/document/d/1b6u-AfjPKfyGZ3orUC7jJUMIJfFUhu3Sk5KOFhjuV0g/edit# | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Expected participants | Application developers, EHR vendors, Business and Data Analysts/Scientists | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Zulip stream | |||||||||||||||||||||||||||||||||||||||||||||||||||||||
Track Kick Off Call | https://us02web.zoom.us/j/85646519661 04/28/2022 10 am - noon Eastern Track Kick-off and Orientation - brief overview on operations and orientation to Track, scenarios and CG IG Discuss resources available Solicit additional scenarios - 'this is your connectathon, what would you like to test out that's not covered?' Follow-up discussions on Zulip thread or confluence. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Agenda |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
Track Details | System roles:
The following is a call-out to the specific operations which will are included in the track. Clients should familiarize themselves with the operations (how to use them and what to expect as a return). Clients are asked to estimate or track their development time. Bonus points for a nice visualization of the parsed and received data. If there is a specific operation that you would like to use for a specific use case or try out, please check with the track leads and they will ensure that relevant data is in the reference implementation. You can send data for inclusion in multiple formats including FHIR or VCF. Servers should review the operations and determine which operations they will be support. The track leads are available for consult and the reference implementation has an accompanying GIT hub. Please have your server prepared 2 weeks before the connectathon. Using the zulip chat you can connect with client participants and track leads maybe able to help with populating data to meet test use case.
If you have a client, server or reference implementation to share you can add it to: http://conman.clinfhir.com/?event=con30 Scenarios These scenarios are examples of using the operation to perform tasks. In the last scenario, we provide some additional examples of use cases for operations. The goal is to test out the operations and gain familiarity with them. Please feel free to work on any additional scenarios that make use of the operations. We hope to gain feedback on utility, completeness and ways to improve or extend the operations! Security and Privacy Considerations: No security or privacy considerations for use of the reference server. The data is synthetic. If implementing a server for Genomics Operations be sure to use test patient data and not expose real patient information. Scenario #0 Name "Variant Normalization - encapsulation through operations" Scenario brief description "Patient has a variant NM_000527.5:c.12G>A in the LDLR gene (see https://www.ncbi.nlm.nih.gov/clinvar/variation/237860/). Search for this variant in patient m123, using the find-subject-specific-variants operation. Modify your search to use any of the variant synonyms on the Clinvar page, to see that the variant is still identified and returned. " Variants can overlap or be subsumed by other variants. Also, optionality in the Clinical Genomics FHIR Variant Profile allows an implementor to represent variants in different ways. As an example the following are synonymous: component:coding-hgvs: HGVS = NM_001195798.2:c.12G>A
component:genomic-ref-seq: NC_000019.10 To be complete, a FHIR query would need to comprehensively include all the different HGVS for a variant, the VCF-like use of multiple components, and perhaps even a list of variants that contain the variant of interest as input parameters. The operations used here in this Track all use a single format for input parameters and guarantee that the search will cover all the possible HGVS and other formatting options. Implicit in these FHIR Genomics operations is that variants in requested regions are returned regardless of how they are formatted/represented/stored in a server. Please read: https://build.fhir.org/ig/HL7/genomics-reporting/branches/operations/operations.html#expectation-of-normalized-variant-search for more information. 'Liftover' is also performed by the server providing the operations. Here is a the base-URL for a public FHIR server that has a result for each of the above possible formats: 'https://api.logicahealth.org/FHIRgenomics2Test/open/'. Patient m123 has a variant profile for each format. But, there are additional patients as well. See if you can find the variants for m123. The FHIR operations for population queries normalizes this with find-population-specific-variants, making finding all the patients with a variant easier. See if you can find all the patient results in the FHIR server with base-URL 'https://api.logicahealth.org/FHIRgenomics2Test/open/' using FHIR queries to learn why operations make looking for data easier. Also, the FHIR operation normalizes the output so you don't need to parse all the various formats, which makes using the data easier. Great news for an app developer, analyst or data engineer. Action (Client): Step 1. Perform find-subject-specific-variants operation to retrieve data. Step 2. Perform the operation again, using a different synonym of the variant. Repeat using several synonyms. Action (Server): Step 1. Populate server with data. Step 2. Implement and provide access to the operation. Precondition: Success Criteria: Operations functional and data is in server. Client is prepared to send REST call and receive data. (end-user UI visualization get's you bonus points). Success Criteria: Data is retrievable and retrieved. Please share operation used, screenshot of result, development time, time for transaction. (Client) Please share server trace log of the transaction and transaction time. (Server) Bonus point: Share screen shot of visualization of result. Comments and critique of the operation. TestScript(s): N/A Scenario #1 Name "Recent knowledge base update" Scenario brief description "Over the course of the past two years, many patients have been tested for variants pathogenic for hereditary breast and ovarian cancer syndrome (HBOC), Lynch syndrome (LS), and familial hypercholesterolemia (FH). Recent knowledge base updates have added to the list of known pathogenic variants, and have reclassified the pathogenicity of many variants. We now want to find patients that have pathogenic variants for HBOC, LS, or FH, based on the new knowledge base." Action (Client): Step 1. Perform find-population-dx-implications operation to retrieve data. Action (Server): Step 1. Populate server with data. Step 2. Implement and provide access to the operation find-population-dx-implications Precondition: Success Criteria: Operations functional and data is in server. Client is prepared to send REST call and receive data. (end-user UI visualization get's you bonus points). Success Criteria: Data is retrievable and retrieved. Please share operation used, screenshot of result, development time, time for transaction. (Client) Please share server trace log of the transaction and transaction time. (Server) Bonus point: Share screen shot of visualization of result. Comments and critique of the operation. TestScript(s): N/A Scenario #2 Name "Population find Cohort matching clinical trial study criteria in genomics" "A researcher has developed a new drug, designed for cancer patients with large deletions involving all or part of BRCA1 (NC_000017.11:43044294-43125364) or BRCA2 (NC_000013.11:32315507-32400268), and wants a list of potential clinical trial participants." Action (Client): Step 1. Perform find-population-structural-intersecting-variants operation to retrieve data. Action (Server): Step 1. Populate server with data. Step 2. Implement and provide access to the operation find-population-structural-intersecting-variants Precondition: Success Criteria: Operations functional and data is in server. Client is prepared to send REST call and receive data. (end-user UI visualization get's you bonus points). Success Criteria: Data is retrievable and retrieved. Please share operation used, screenshot of result, development time, time for transaction. (Client) Please share server trace log of the transaction and transaction time. (Server) Bonus point: Share screen shot of visualization of result. Comments and critique of the operation. TestScript(s): N/A Scenario #3 Name "HLA matching and implications" "A patient has undergone HLA typing, and clinicians wants to know if there is any haplotype or genotype indication that HLA-A*23 is present". Action (Client): Step 1. Perform find-subject-specific-haplotypes operation to retrieve data. Action (Server): Step 1. Populate server with data. Step 2. Implement and provide access to the operation Precondition: Success Criteria: Operations functional and data is in server. Client is prepared to send REST call and receive data. (end-user UI visualization get's you bonus points). Success Criteria: Data is retrievable and retrieved. Please share operation used, screenshot of result, development time, time for transaction. (Client) Please share server trace log of the transaction and transaction time. (Server) Bonus point: Share screen shot of visualization of result. Comments and critique of the operation. TestScript(s): N/A Scenario #4 Name "Genetic related diagnosis of patients like mine" "Physicians have wondered 'What do patients with my patient's variants have in common with regards to reported genetics-linked diagnostic implications?' In this scenario, use a set of variants to represent your physician's patient's variants to find a population in the server that has those variants (or various subsets). Then, use that population and find the implications associated with the variants found in the population. Try a smaller list of variants first. Be creative." Action (Client): *suggested steps, be creative and use the operations and process that make the most sense to you Step 1. Perform find-population-specific-variants, and/or find-population-structural-intersecting-variants, and/or find-population-structural-subsuming-variants operation, as appropriate, to retrieve list of patients. Step 2. Iteratively perform find-subject-dx-implications. Step 3. Compare the diagnostic implications of the population to the patient. Action (Server): Step 1. Populate server with data. Step 2. Implement and provide access to the operations to meet the use case. Precondition: Success Criteria: Operations functional and data is in server. Client is prepared to send REST call and receive data. (end-user UI visualization get's you bonus points). Success Criteria: Data is retrievable and retrieved. Please share operation used, screenshot of result, development time, time for transaction. (Client) Please share server trace log of the transaction and transaction time. (Server) Bonus point: Share screen shot of visualization of result. Comments and critique of the operation. TestScript(s): N/A Scenario #5 Name "Additional Examples" Feel free to devise additional scenarios to test. Here are some explicit examples:
Action (Client): Step 1. Perform operations to retrieve data and any additional logic for the use case. Action (Server): Step 1. Populate server with data. Step 2. Implement and provide access to the operation Precondition: Success Criteria: Operations functional and data is in server. Client is prepared to send REST call and receive data. (end-user UI visualization get's you bonus points). Success Criteria: Data is retrievable and retrieved. Please share operation used, screenshot of result, development time, time for transaction. (Client) Please share server trace log of the transaction and transaction time. (Server) Bonus point: Share screen shot of visualization of result. Comments and critique of the operation. TestScript(s): N/A Security and Privacy Considerations: No security or privacy considerations for use of the reference server. The data is synthetic. If implementing a server for Genomics Operations be sure to use test patient data and not expose real patient information. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||