August 2023 Clustering User Interface Improvement
While optimizing CODI during its beta phase, we noticed some inconsistencies in the clustering widget when using the DBSCAN* algorithm, whereby the CODI user interface misrepresents information pertaining to number of clusters and localizations.
Data CSVs exported from CODI were not affected and output the correct clustering and localization information.
As of August 24, 2023, CODI was updated to display the correct information. As a result, you may notice some differences in the number of localizations and clusters in the following tools: Clustering and Counting.
How do I know if I am impacted?
This issue affects CODI users who use the DBSCAN* clustering algorithm in either the Legacy, Clustering, or Clustering and Counting Apps, and rely on the CODI user interface for quantifying results.
Note: This does not affect clustering results in the EV Profiling App
What is the impact on your data?:
Note: Data CSVs output by CODI were not affected and continue to output the correct clustering and localization information.
Certain datasets are more affected than others; the CODI team recommends you evaluate the impact by comparing to cluster / localization counts that you had previously logged.
What you need to know:
- Number of clusters
- Number of clustered localizations
- Number of localizations within a selected cluster ID
We were able to identify this discrepency by:
- Using the "download clusters CSV" button in the "..." menu in the top right of the Clustering tool, in any of the CODI Apps or Legacy analysis.
- Sum the total number of localizations for clusters in the selected channel; the total number of clustered localizations differed from the value displayed in CODI
- NOTE: For multi-channel datasets, ensure you are comparing values to the actively selected clustering channel in the CODI user interface (right side at the top of the channels list, with the clustering icon)
- Identifying a single cluster from the CSV by its Cluster ID, entering this Cluster ID value into the "selected cluster ID" input in the "..." menu in the top right of the Clustering tool. The displayed number of "locs" differed slightly from that in the exported CSV file.
Other related changes:
- The "download clusters CSV" button now works reliably from the "..." menu of the Clustering tool in any of the CODI Apps or Legacy analysis.
We apologize for any inconvenience this issue may have caused and thank you for your continued support and patience. If you have any questions or need further assistance, please don't hesitate to reach out to our support team.
Thank you,
ONI Team
Related Articles
How to guide: Clustering tool
Tool view Purpose of tool The tool performs clustering on localisation data to find biological structures and quantify their statistical and morphological properties. The default clustering method is HDBSCAN. The cluster results can then be ...Release Note: CODI Apps and Batch Analysis - June 2023 Release [Updated August 2023]
Welcome to the latest release notes for CODI Apps and Batch Analysis! For questions about the new software release, please contact our Customer Success team on https://help.codi.bio/ or any software team member directly via ONI-Club slack channel. ...How to guide: Counting tool
Tool view Purpose of tool Counts the number of localisations or ‘bins’ by channel around a set of centre points - currently the centroids of a clustering result. Contents Inputs Results Outputs Inputs The centre points for counting come from a loaded ...CODI-FAQ
What is CODI ? Our CODI cloud platform is a beta and is currently an analysis platform used for the analysis of data acquired by our Nanoimager and its local acquisition software (NimOS). CODI is... A collaborative platform to share your data and ...How to guide: The CODI Desktop Uploader
Tool view Purpose of tool The CODI Desktop Uploader is a tool created to ease the transition between NimOS and CODI. The main objective of the application is to facilitate the uploading of NimOS generated datasets to the CODI cloud platform. It ...