Thanks for your interest in leading an OMIQ training or data analysis course!
This guide will give you some guidelines and best practices for running an interactive OMIQ session.
Let the OMIQ Team know you’re running an event
We’re happy to enable 3rd party or institutional level training and the more we know, the more we can help you out.
Please reach out to support@omiq.ai and let us know the date, time, and relative number of users; however we don’t require a final attendee list as we can provision a dedicated training code for anyone interested in attending. The OMIQ support team can create this special Workshop license code with a flexible duration to enable seamless account access for attendees during the event and afterwards. This workshop license won’t show up on their profile/account details page, so please be clear with attendees about the pre-arranged duration. By default a free OMIQ trial is 4 weeks, however if you need more time for a particular class or training effort we can oblige on a case-by-case basis. You can include the following instructions to the users to claim this code
To claim the OMIQ Workshop trial license, please go to app.omiq.ai (or your institutional server) and if you don’t yet have one, create an account using your institutional email. After creating an account, you’ll get a validation email to make sure the email was entered correctly - please click on the link provided therein and then log into OMIQ. Open the Profile tab at the top of the page, click on the menu next to the Managed Seats section, and then click Claim License Key. Enter the provided Workshop license code there and click Update.
Depending on the email domains that people will be registering with, we may need to approve them individually. Please confirm that everyone interested in attending the training has an active OMIQ account at least 3 days before the training is scheduled to start so we have enough time to sort out any account issues before the day of the training.
Class participants should share a single dataset, not upload their own
This is one of the only rules we ask you to abide by, as it will ensure the stability and efficiency of the OMIQ servers for the duration of the training. If each participant uploads their own copy of a dataset, even if the files are technically the same, OMIQ will treat them as unique copies and efficiency will suffer during the training. When everyone accesses the same Dataset in OMIQ, even if they are using different Workflows for their analyses, OMIQ will reference the same underlying files and will be much more responsive.
–Again, do not have your users each upload unique data for the Workshop–
As the organizer, you should upload a single copy of the files and then share with the attendees. They can each still make their own Workflows within that dataset.
If this course is part of a larger effort that includes each user or smaller group acquiring and then analyzing their own files, we still ask that the training be performed on a single dataset. Besides efficiency benefits in OMIQ, analysis principles are best demonstrated on known, high-quality data and we’ve seen a large variance in the quality of files produced in these settings based on both personal and professional experience. Once out of the workshop setting they’re certainly free to analyze their own/group’s data using the strategies demonstrated.
Collaborate using Groups, not individual connections
As demonstrated in the OMIQ Tutorial, there are two ways to collaborate in OMIQ. You can connect with individual named users via email or you can add a Group which is a collection of users with a set permission level.
Using a Group makes it easier to get users access to the resources they need and keeps you from having to manage lists of individual users for each resource.
All Workshop participants should be added to the Group. You can then share the core training Dataset and pre-made example Workflow with the Group and set their permissions level to Reader.
To make a Group, go to your Profile page and near the bottom click on the Create Group button.
You can then invite individual pre-existing Connections or you can have Workshop participants add themselves via the Join Group by Code.
Use the Group Join Code to add users to the Group
To streamline the Group building process, you can create a Join Code after clicking on the new Group you’ve just created.
This will generate a string of characters that each Workshop participant can enter on their Profile page under the Join Group by Code option. We’ve found it useful to both display the code on the screen but also type it into any chat window present as part of the training so users can copy and paste the code directly. This allows users to add themselves to a group and is especially useful as class size scales.
After a Group is given access to a Dataset, you can continue to add additional users to the Group or add further datasets to the group at your discretion.
Final Tips
We’ve found it to be very helpful to have both the Dataset and an established Workflow completed before the training starts so that if any participants run into issues or need to catch up to the group, they can copy the already finished workflow. This also helps to prevent time intensive tasks from derailing the flow of the training.
If performing the class in person and resources allow, it can be helpful to have one person leading the training and another person walking around the room to help individuals who have questions or get off-track.
It’s best to have the example dataset that’s complex enough to demonstrate tasks of interest, but not too complex that new users lack the ability to understand the scope of the experiment. Let us know if you’d like a recommended one!
When running the example Workflow, keep track of timing on algorithm tasks to make sure that each participants’ tasks will be able to complete in the time allotted. Consider using fewer events via the Subsampling task, or faster algorithms (e.g. opt-SNE instead of tSNE). For normal PBMC Phenotyping analysis anywhere from 10k-50k events is a good starting point.