Skip to content

OME-Zarr troubleshooting

OME-Zarr is a developing standard and is very flexible, which sometimes makes it hard to read and write, and no software supports all features.

Microscopy Nodes supports OME-Zarr up to version 0.5, to load single, up to 5-dimensional, arrays.

Is your OME-Zarr not loading?

A quick option is to append /0 after your path. Some OME-Zarr writers create a group at the .zarr adress, with the first (and often, only) image at .zarr/0

OME-Zarr collections

There is currently no support for any form of self-discovering collections from OME-Zarr, this can cause issues for:

  • Wells
  • Fields
  • Labels
  • Large Zarr-groups (such as OpenOrganelle datasets)
  • Bioformats2raw export

All of these images can still be opened, by pointing to the specific path of the array, which contains the different OME 'multiscales'.

Pointing to a specific path

For example, for the dataset https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.3/idr0052A/5514375.zarr of the IDR, the labels are at https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.3/idr0052A/5514375.zarr/labels/Cell and https://uk1s3.embassy.ebi.ac.uk/idr/zarr/v0.3/idr0052A/5514375.zarr/labels/Chromosomes. These are not auto-discovered, but can be loaded with the specific paths.

Support for collections will come in the future, but this will wait for the planned reorganization OME-Zarr collection structure.