Our engineering team discovered an issue in the randomization tool algorithm that assigns animals to study groups. Between Dec 18, 2023 and June 5, 2024, study groups using the clustered randomization option were potentially less optimized than expected.
When allocating animals to groups based on more than one metric, animal measurement data is used to assign animals to groups by applying an algorithm called “kMeans clustering”. This algorithm had been impacted insofar as the outcome of the randomization had resulted in less than optimal distribution of animals across treatment groups.
This randomization feature allows the user to detect suboptimal randomization by reviewing the group summary data (Mean, SD/SEM, Median,MAD), and ANOVA calculated P-values and correcting this prior to submitting randomization. Users can also manually correct group assignments at any time.
Benchling expects users to review the group summary to ensure that the randomized groupings are sufficient for the study needs. Randomization reports can be viewed for each study in the Randomizations tab of the Overview section.
If you have questions or would like more information, please reach out to support@benchling.com.