Skip to content

samplesheet.md

Latest commit

 

History

History
75 lines (58 loc) · 3.88 KB

samplesheet.md

File metadata and controls

75 lines (58 loc) · 3.88 KB

Samplesheet format

The samplesheet is a CSV with one row per pair of adjacent sections (one MERSCOPE, one Xenium per pair). A template lives at workflows/samplesheet.example.csv.

Columns

Column Required Description
pair_id yes Unique identifier for this sample pair. Used as the top-level output directory name.
merscope_dir one of these two Path to the raw MERSCOPE region export folder (contains transcripts.parquet, cell_boundaries/, images/, etc.).
merscope_spatialdata_path one of these two Path to an existing (or desired) reusable MERSCOPE SpatialData zarr. If it exists, the build step is skipped unless --force_spatialdata_build true is passed to Nextflow.
merscope_image_prefix no Prefix used to match z-plane image keys when more than one run is present.
merscope_z_range no Inclusive z-layer range as start-end. Defaults to 0-6.
merscope_transform_path no Override path to the micron_to_mosaic_pixel_transform.csv. If not set, MerXen looks inside merscope_dir.
merscope_channels no Comma-separated channel names for Cellpose. Defaults to DAPI,PolyT.
merscope_voxel_layers no ProSeg voxel layer count for MERSCOPE. Defaults to 7 (from nextflow.config).
xenium_dir one of these two Path to the raw Xenium export folder.
xenium_spatialdata_path one of these two Path to an existing (or desired) reusable Xenium SpatialData zarr. Cached the same way as the MERSCOPE path.
xenium_channels no Comma-separated channel names for Cellpose. Defaults to DAPI,18S.
xenium_min_qv no Minimum transcript quality value to retain. Defaults to 20.
xenium_voxel_layers no ProSeg voxel layer count for Xenium. Defaults to 2.
xenium_spec_path no Override path to experiment.xenium or specs.json used to derive the micron→pixel transform.

Aliases

merscope_zarr_path is accepted as an alias for merscope_spatialdata_path for backwards compatibility.

Validation rules

From workflows/main.nf:71-90:

  • Every row must have a non-empty pair_id.
  • Every row must provide either merscope_dir or merscope_spatialdata_path.
  • Every row must provide either xenium_dir or xenium_spatialdata_path.

The Python-side parser lives in src/merxen/io/samplesheet.py:52. It is used by unit tests and can be invoked directly by scripts, but the Nextflow workflow parses the CSV itself.

Minimal example

pair_id,merscope_dir,xenium_dir
EXAMPLE01,/path/to/merscope/EXAMPLE01/region_R1,/path/to/xenium/EXAMPLE01

This is enough to run the whole pipeline end-to-end with default parameters. All other columns fall back to defaults.

Full example

pair_id,merscope_dir,merscope_spatialdata_path,merscope_image_prefix,merscope_z_range,merscope_transform_path,merscope_channels,xenium_dir,xenium_spatialdata_path,xenium_channels,xenium_min_qv,merscope_voxel_layers,xenium_voxel_layers,xenium_spec_path
EXAMPLE01,/path/to/merscope/EXAMPLE01,/path/to/cache/EXAMPLE01_merscope.zarr,,0-6,,"DAPI,PolyT",/path/to/xenium/EXAMPLE01,/path/to/cache/EXAMPLE01_xenium.zarr,"DAPI,18S",20,7,2,
EXAMPLE02,/path/to/merscope/EXAMPLE02,,,,,"DAPI,PolyT",/path/to/xenium/EXAMPLE02,,"DAPI,18S",20,7,2,

Row 1 uses cached SpatialData zarrs if they already exist. Row 2 builds everything fresh from the raw exports.

Tips

  • Quote any list-valued field that contains commas (merscope_channels, xenium_channels).
  • Leave a field empty (,,) to fall back to the Nextflow default.
  • Prefer absolute paths. Relative paths are resolved from the shell that ran nextflow, not from the work directory.
  • Keep one samplesheet per batch. Nextflow runs all rows in parallel up to the executor limit.