Skip to content

Latest commit

 

History

History
117 lines (67 loc) · 11.5 KB

training.rst

File metadata and controls

117 lines (67 loc) · 11.5 KB

Prepare an OMERO server for training

This chapter describes a setup of a training OMERO.server at your institution. Further, it gives an overview of the other training resources as well as hints for trainers about how to present OMERO during a training session.

Resources

Download of scripts and OMERO.cli environment setup

  • This guide assumes that you have installed an OMERO.server for training purposes.

  • Clone the training-scripts repository:

    $ git clone https://github.com/ome/training-scripts.git

  • Set up a OMERO.cli as specified under CLI installation. Typically, this environment will be used on your local machine. Alternatively, you can use the OMERO.cli environment of the OMERO.server.

Setup of groups and users

Prepare groups and users as listed in the provided template sheet. Assuming you have the template sheet in the same directory as the script create_groups_users.sh, you can run:

$ cd training-scripts/maintenance/scripts
$ HOST=$YOUR_SERVER_ADDRESS PASSWORD=$PASSWORD_FOR_ROOT bash create_groups_users.sh

which will create 50 users in your database. The users are members of four main OMERO.groups, which cover the group permissions allowed in OMERO. The Read-annotate group Lab 1 is the main group used in the trainings, which contains most of the data in a typical training. The setup in the template sheet sets this Lab 1 group as a default group in OMERO for all users.

Rename users to have first and last names of real people (the list of famous scientist names is used) running the renaming script:

$ python rename_users.py trainer-1 $PASSWORD --server $YOUR_SERVER_ADDRESS

Import images

We like to use in-place import because it avoids duplicating the data on disk, but if you are only importing a small amount of data you may find non-in-place import (described below) is more convenient. In-place import is achieved by using import bash script. Use the publicly available data provided or use your own data or data downloaded from IDR. Shell into the machine where you have installed your OMERO.server (necessary for in-place import) and, assuming that you have for example your data in a folder my-folder which is visible from that machine, run:

$ HOST=localhost SUDOER=trainer-1 PASSWORD=$PASSWORD_FOR_trainer-1 IMPORTTYPE=normal NUMBER=15 FOLDER=/path/to/my-folder bash in_place_import_as.sh

This will in-place import the images from my-folder into a new Dataset named /path/to/my-folder for user-1 through user-15.

To achieve a better reproducibility of your imports, you can use a bulk file pointing to a list of paths of the image files. Assuming you have for example the bulk file idr0021-experimentA-bulk.yml located on the filesystem you are working on, together with the corresponding idr0021-experimentA-filePaths.tsv file and the images as specified in the idr0021-experimentA-filePaths.tsv which you have edited accordingly, you can run:

$ HOST=localhost SUDOER=trainer-1 PASSWORD=$PASSWORD_FOR_trainer-1 OMEUSER=trainer NUMBER=2 BULKFILE=/path/to/idr0021-experimentA-bulk.yml bash in_place_import_as.sh

to import the images for trainer-1 and trainer-2.

The script assumes by default in-place imports. You can adjust it for "classsical", non-in-place import by deleting the --transfer=ln_s from the script lines or, if using the bulk file workflow, comment out the transfer line from the bulk file.

Image data can also be populated by a Python script which copies images as pixeldata (i.e. not the original images) from IDR by default. You can adjust the three lines beginning with idr_client = omero.client(host="idr.openmicroscopy.org", port=4064) to copy from other OMERO.servers. Note that the script is only creating images of single Z and T, and thus reducing the original images dimensions in case these are multi-Z or T images.:

$ python idr_copy_plate.py trainer-1 --server $YOUR_SERVER_ADDRESS $PASSWORD $PLATE_ID

This will copy the Plate with $PLATE_ID from IDR as trainer-1 into your server, but note that this idr_copy_plate.py script only works on Plates which have a single Field (Image) per Well.

Import metadata

Import metadata from a CSV via OMERO.web or Command Line Interface.

Annotate images using training-scripts:

  • Copy Key-Value pairs from IDR or other OMERO server. This script will copy MapAnnotations from Images in IDR to Images with the same names in your server (Datasets names must also match). The command below copies from Project:1 in IDR to Project:2 in your server. The Project in your server is owned and the action is executed by trainer-1.:

    $ python idr_get_map_annotation.py trainer-1 $PASSWORD_FOR_trainer-1 Project:1 Project:2 --server $YOUR_SERVER_ADDRESS

  • Add new Key-Value pairs. The command below will add Key-Value pairs defined inside the script randomly to the images inside Datasets with name big-dataset for all 50 users in your server. The $PASSWORD for all the users must be the same.:

    $ python key_value_pairs.py $PASSWORD big-dataset --server $YOUR_SERVER_ADDRESS

  • Calibrate images. The command below will calibrate all the images within Dataset named western-blots to 0.33 micrometers per pixel for all 50 users in your server. The $PASSWORD for all the users must be the same.:

    $ python calibrate_images.py $PASSWORD western-blots --server $YOUR_SERVER_ADDRESS

  • Add timestamps. The command below will set timestamps on the timelapse images within Dataset named timestamps with delta T of 300 seconds for all 50 users in your server. The $PASSWORD for all the users must be the same.:

    $ python set_timestamps.py $PASSWORD timestamp --server $YOUR_SERVER_ADDRESS

  • Propagate tags and ratings to all users. Supposing that the trainer-1 has a Dataset to-tag with Tags and Ratings on the images in the Dataset. Further, each user, such as user-1, user-2 has the same-named Dataset with equivalent images in it, but with no Tags and Ratings (a typical situation after a fresh import of images). The command below will link the Tags of trainer-1 which are linked to the images in the to-tag Dataset to the corresponding images in the to-tag Datasets of the users. The links between the Tags and the Images will belong to each user. Also, the Ratings which are on the Images of the to-tag Dataset of trainer-1 will be re-created for the corresponding Images of the users and will belong to those users.:

    $ python copy_tags_ratings.py to-tag $PASSWORD --server $YOUR_SERVER_ADDRESS

Add analytical metadata

Create an analysis results table using a script run from a 3rd party tool. For example, you can run the segmentation script in the scripting editor of Fiji on a Project in OMERO containing Datasets with Images which creates an OMERO.table and a CSV file with results and attaches these to that Project in OMERO.

These analytical results can be used to showcase OMERO.parade.

Cleanup scripts

It might be of great advantage to be able to clean up in batches, but still selectively, metadata added to the images on your training server.

  • Delete ROIs on all Images inside Datasets of specified name for all users on the server who have such Datasets. In the example below, the Dataset's name is with-rois. The $PASSWORD is the password of the user deleting the ROIs. The deleting user is trainer-1 by default.:

    $ python delete_ROIs.py --datasetname with-rois --server $YOUR_SERVER_ADDRESS $PASSWORD

  • Delete Annotations on all Images inside Datasets of specified name for all users on the server who have such Datasets. In the example below, the Dataset's name is western-blots and the deleted annotation type is FileAnnotation. All other annotation types such as TagAnnotation etc. will be preserved. The $PASSWORD is the password of the user deleting the ROIs. The deleting user is trainer-1 by default.:

    $ python delete_annotations.py --anntype file  --namespace none --server $YOUR_SERVER_ADDRESS $PASSWORD western-blots