Skip to content

About the GKM Starter Kit

The GKM Starter Kit is a community project for sharing and using genomic knowledge represented with the Genomic Knowledge Model.

What is GKM?

The Genomic Knowledge Model (GKM) defines shared models for genomic variation and related knowledge. It is part of the broader GA4GH Genomic Knowledge Standards (GKS) Work Stream.

GKM brings together four standards that can be used independently or combined:

  • GKM-Core provides concepts shared by the other standards.
  • VRS describes precise molecular variation.
  • Cat-VRS describes categories of variation.
  • VA-Spec describes what is known or asserted about variation, including supporting evidence.

Producers use the standards that fit their data. Their Python reference implementations are:

The three pillars

The Starter Kit connects GKM standards to real data, software, and community outcomes through three pillars.

I. Data

Data bundles package related GKM objects while preserving the names and organization of their resources.

A documented, versioned bundle schema makes structural changes easier to communicate and helps consumers understand what a producer distributes.

Bundles are optional in GKM

A GKM bundle packages related GKM objects using a producer-defined structure. Bundles are not required to use GKM: applications can construct, validate, and serialize individual GKM objects through the reference implementations above.

II. Tools

ga4gh.gkm provides a consistent Python interface for loading, exploring, and writing bundles.

Consumers can work with supported GKM Python models, follow relationships between objects, and retain producer-specific content without hand-parsing each producer's organization.

Bundles are the input and output format of ga4gh.gkm.bundles. The initial implementation supports JSON serialization; other serializations may be added later.

The GKM reference implementations construct, validate, and serialize individual GKM objects; ga4gh.gkm.bundles works with collections of those objects in bundles. These packages are maintained implementations, not required interfaces. Developers can implement the GKM standards in their own tools and programming languages.

III. User Stories

User stories show how communities put GKM standards to work and what their implementations accomplish.

They give prospective adopters concrete examples, reveal reusable patterns, and help teams evaluate whether GKM fits their needs. Each is told as a short vignette drawn from a real implementation.

Some user stories use data bundles or ga4gh.gkm; others use GKM standards without them. The focus is the real-world outcome, not whether an implementation uses every part of the Starter Kit.

How they work together

Data bundles provide real, packaged knowledge, ga4gh.gkm provides a consistent bundle workflow, and user stories demonstrate the results. Together they help producers, consumers, and the GKM community test the standards, identify gaps, and turn successful implementations into patterns others can reuse.