Zarr-based format for storing quantitative biosystems dynamics data

The following is some progress of storing ROIs and tables for tracking

These are just test codes to test out ideas

https://docs.google.com/presentation/d/1_HGms52mzrZMWRSIeevzlaQ-QS48QBGp-hLFvObKBQo

Some description from the Global BioImaging Hackathon:

Our initial goal was to prepare some kind of container to store the coordinates of biological objects such as the positions of molecules and nuclear contours (and their tracking information) detected by image processing. However, if, for example, the nuclear contour is represented as a set of vertices, it is difficult to store them in Zarr because of their variable number. Knowing that there is a framework of AnnData, we think of using OME-Zarr Mask to describe the ROI based on pixels associated with the image, and storing the representative position for biological object in each row of X in AnnData. We think that the features for each biological object such as area, volume, mean value of GFP signal, etc., can be stored in obs. We think the same issue arises in spatial transcriptome data when describing data for a cell that are detected across multiple pixels.

We think it would be hard to understand without some exmaples, so we would like to create some examples.

We have already developed several data formats based on XML and HDF5, but we want to make a new format highly compatible(?) with OME-NGFF.

• https://doi.org/10.1371/journal.pone.0237468
• https://doi.org/10.1093/bioinformatics/btu767

• https://so.qbic.riken.jp/ssbd/zarr/v0.3/wt-N2-081015-01.ome.zarr/
• https://uk1s3.embassy.ebi.ac.uk/bdz/0.1/wt-N2-081015-01.ome.zarr (updated 22 April 22)
  • Special acknowledgement to EMBL-EBI’s Embassy Cloud and the BioImage Archive for providing valuable S3 storage and support for this project.
• https://so.qbic.riken.jp/ssbd/zarr/v0.3/wt-N2-081015-01.ome.zarr.zip

Kyoda, K., Okada, H., Itoga, H. and Onami, S.: ‘Deep Collection of Quantitative Nuclear Division Dynamics Data in RNAi-Treated Caenorhabditis Elegans Embryos’. bioRxiv, https://doi.org/10.1101/2020.10.04.325761.

It requires the module from Kevin Yamauchi:

https://github.com/kevinyamauchi/ome-ngff-tables-prototype

pip install git+https://github.com/kevinyamauchi/ome-ngff-tables-prototype

https://github.com/openssbd/bdz/blob/main/view_kyoda_wormdata.ipynb

[updated 22 Apr 22 - currently there is error on accessing tracking data via S3 ... under investigation]