Skip to content

inputs.md

Latest commit

 

History

History
91 lines (59 loc) · 1.85 KB

inputs.md

File metadata and controls

91 lines (59 loc) · 1.85 KB
Raw

Input Data

Baysor consumes a molecule table plus an optional prior segmentation input.

Supported Molecule Inputs

The coordinates positional argument may be:

  • CSV molecule table
  • Parquet molecule table
  • experiment.xenium

When experiment.xenium is used, Baysor resolves the adjacent Xenium transcript table automatically and keeps enough source context to make the legacy outputs compatible with xeniumranger import-segmentation.

Required Molecule Columns

At minimum, Baysor needs:

  • x
  • y
  • gene

If the data are 3D, it can also use:

  • z

Column names are configurable through CLI flags or the config file.

Common Column Mapping Flags

  • --x-column
  • --y-column
  • --z-column
  • --gene-column
  • --qv-column

For Xenium, configs/xenium.toml already maps:

  • x_location
  • y_location
  • z_location
  • feature_name
  • qv

Optional Source Columns

Depending on the workflow, the input table may also carry:

  • a transcript-native prior column such as cell_id
  • confidence / quality columns
  • Xenium transcript_id

For Xenium-origin inputs, Baysor preserves transcript_id so it can be written back to legacy output when appropriate.

Cropping During Input Load

The C++ branch supports spatial filtering during input loading:

  • --x-min, --x-max
  • --y-min, --y-max
  • --z-min, --z-max

This is useful for:

  • quick development runs
  • protocol debugging
  • visualization crops
  • testing large datasets without loading the full field of view

Protocol Notes

Xenium

Preferred input:

./build/baysor run -c configs/xenium.toml data/experiment.xenium :cell_id

ISS / osm-FISH / STARmap

These usually use CSV molecule tables plus either:

  • no prior, with explicit --scale
  • an image mask prior
  • a boundary file prior

See Examples for runnable commands.