Format Specification

Overview

The Spacewalk binary file format (.sw) uses HDF5 to store and organize genomic spatial data. The format supports both single-point (centroid) and multi-point (cluster) data types.

File Structure

Header Group

Required top-level group named header with metadata:

Key Value
format “sw”
genome A valid genimd id. e.g., “hg19”
pointtype “SINGLE_POINT” or “MULTI_POINT”

Genomic Position

The genomic_position group contains:

  • regions dataset: 2D array with columns:
    • chromosome name (string)
    • region start (integer, bp)
    • region end (integer, bp)

Regions must be sorted by genomic start position.

Spatial Position

The spatial_position group contains XYZ data in two possible formats:

Single Point Format

  • One XYZ point per genomic region
  • Datasets named t_0, t_1, etc. for each trace
  • Direct 1:1 mapping between regions and points

Single Point Association

Multi Point Format

  • Multiple XYZ points per genomic region
  • Trace groups named t_0, t_1, etc.
  • Each row format: region_index x y z
  • Flexible mapping between regions and points

Example Files

View example files in HDF5 format: