Curation Projects - In Detail
Gwenn Berry avatar
Written by Gwenn Berry
Updated over a week ago

Curation projects allow you to collaborate with your team or organize your own work to curate towards a goal. By splitting curation tasks into manageable and focused projects, and across team members, you can make incremental and measurable progress towards a higher level of data quality and avoid curation fatigue.

There is no "right" size or goal for a curation project, but we recommend batching into projects that can be completed within a couple weeks' time, and with concrete goals. For example, you might choose a goal similar to one of these:

  • Curate all SNVs over 1% VAF for all non-cancer datasources

  • Curate 100% of results for 20 selected datasources

  • Review each of 20 selected datasources for "low-hanging fruit" results to curate, stopping when you have curated 50 results or have completed review for all datasources

Create and Configure Your Curation Project

Create Project and Add Sources

  1. From your home screen, go to Curation Projects. Click the Add button at the top right. Give your project a title and description, then click Save.

  2. Add Resultsources to the project for review by clicking Add Sources in the Quick Action toolbar. Select a workflow version and filter data sources by any desired tags. Click Check Sources if you want to sanity check your query, and Add Sources to submit.

  3. You can optionally batch-assign the newly added Resultsources to yourself or a collaborator by using the Assign To... dropdown.

Default Result Filters

Adding default filters allows you to focus on a subset of results in the manual curation process and avoid entering in the same filters for every sample you review. For example, you may choose for a particular project to only focus on the extremes to take care of the "lowest hanging fruit" true and false positives before moving on to the more challenging "gray area" results. To save time and avoid data clutter, apply a default filter to the project that only presents you with these "extreme" results.

When opening a Resultsource for curation from the curation project, the default filters will be automatically applied so that you will only see the subset of data you care about. Of course, you can always clear or modify the filter from there using the filter panel in the Resultsource view, or further filter using the table and column controls.

To set or modify the default result filter, unlock the setting (if applicable) in the Results card of the Curation Project page, add your filters (using the Query Language), and click the button to submit.

Adding Collaborators

In order to assign Resultsources to your teammates for curation, they need to be added as collaborators on your project. Add teammates by clicking the + icon next to the list of collaborators.

Assigning and Removing Resultsources

Resultsources can be assigned, unassigned, or unlinked from your project using the Batch Edit feature. Simply select the rows you want to edit using the checkboxes on the left, then click Batch Edit and choose the relevant action from the modal.

You can also unassign or self-assign any resultsource in the Result Sources table using the buttons in the Assignee column.

Curating and Exploring Your Configuration Project

Curate Resultsources

To begin (or resume) curating a resultsource, first check that it is assigned to you (and self-assign as needed). Open the resultsource by clicking the source's name in the left column. This will open to the standard Resultsource view, but will also attach the curation project information to it. It will also automatically trigger the Resultsource view to open in Curation mode.

It is best to access curations from the project page, opening the Resultsource from outside the project context will cause you to lose the linking information. However, if you have a curation triage that you created outside of the project context, it can still be linked after the fact.

Proceed with curation as described in Performing Resultsource Curation. You may wish to Save your work regularly by creating and updating a saved triage. You can share this triage with collaborators or come back to it at any time, and when accessing the Resultsource from the project context it will automatically open with the saved triage if there is one created.

When you are ready to close out this round of curation for a given Resultsource, you can submit your triage to finalize and apply your curations. This will automatically mark the Resultsource Curation as complete (unless you select Keep Open) and you can then move on to another Resultsource.

You can also manually change the status of a Resultsource Curation (for example, to re-open a mistakenly completed curation, or to flag or defer for future review). Within the Resultsource Curation view, click the status indicator at the top right to drop down a list of possible statuses to set.

View All Results

To view all results linked to a curation project, click View All from the Results card of the Curation Project page. This will show you an aggregated list of all results under review from all sources in the project.

If you have a default results filter on your project, that will automatically be applied as the inclusion filter. You can also use the filters to "flag", rather than include/exclude, results meeting certain criteria (or do a combination). You can then use that flag status (furthest right column), along with any other column filters, to filter/select results in the table and apply batch actions or simply output as annotation.

Secondary flag filters in the View All Results view are not saved as part of your data curation. They are meant to facilitate batch actions, exploration, and annotations with complex or tiered filter conditions.

Did this answer your question?