MAF to Tile CSV Architecture - GenomicsDB/GenomicsSampleAPIs GitHub Wiki

High-level Import Architecture

The data import architecture is layered to provide flexibility and code re-use. It has 3 key pieces - the first 2 are used to generate the CSV and the 3rd is the DB itself.

• Common Translation layer - provides the functionality that all the Input translation layer plug-ins will need. This layer provides the tools, book keeping, checkers, and formatting that ensure that the output CSV file is compatible with the Tile DB loader modules.

• Input Translation Layer - handles the input specific translation and transformation before data can be consumed. This layer has to be written for every new input format. This layer is tightly coupled with the Common Translation layer, and leverages the functionality to convert the given input to tile DB CSV format.

• Variant DB - Is a composite of two databases.

  • Tile DB - is the repository for the variants
  • Meta DB - is the repository that stores the different associations of the data contained in Tile DB in order to support GA4GH