Imaging

Introduction

Imaging data are the images produced by microscopy of cells, organelles, tissues etc and their associated metadata.

OMERO

The Sanger Imaging Platform uses the OMERO software for the analysis of microscope imaging data. Storing data in OMERO is a procedure Cellular Genetics Informatics provide.

Please watch the following 10min video from the The Jackson Laboratory to learn the basics of browsing and annotating images using OMERO:

OMERO uploads

Users need to get data off microscopes quickly. Choose from the following use cases on how to request your data be uploaded to Sanger’s OMERO server.

Upload OME-TIFF and NDPI

Data in OME-TIFF and NDPI format can be imported as-is provided the images have pyramids generated. Please submit a ticket with a TSV file with the following columns:

OMERO_IMPORT.tsv
filename	Name of the file (`my_file.ome.tiff` or `my_file.ndpi`)
location	Path to the file folder (ie: `/nfs/team123/`)
omero_group	OMERO group name (ie. `MJ_CDB`)
omero_project	OMERO project folder name (first level folder)
omero_dataset	OMERO dataset folder name (second level folder)
omero_username	Image owner’s username (ie. `ob1`)

Upload NDPIS

NDPIS need to be processed before importing this involves merging all NDPI image for individual channels into one image. Please fill the following xlsx file and submit a ticket with it attached.

HamamatsuLogTemplate.xlsx

Stitching

Phoenix exported tiles need to be stitched together before importing to OMERO. Please fill the following xlsx file and submit a ticket with it attached.

PhonexLogTemplate.xlsx

Segmentation

We can also run segmentation on your images provided you give us the following information:

input_image	Full path to the image file. ie. `/nfs/team123/my_file.tiff`
dapi_ch	DAPI channel index (default `0`)
cyto_ch	Cyto channel index (default `''`)
object_diameter	Object diameter (default `35`)
flow_threshold	Flow threshold (default `0.5`)
expand	Cell border expansion ammount (default `35`)
magnification	Magnification level (default `20`)
model_type	Cellpose model type (default `cyto`)

Our segmentation pipeline works on GPU using Cellpose in order to produce 2D or 3D images of cells and nuclei with segmentation for downstream analysis.

If you want to run the pipeline yourslef you can do so following the instructions on GitLab https://gitlab.internal.sanger.ac.uk/cellgeni/imaging/segmentation-cellpose

Registration

TBD