Model spots outlier control groups in mouse lifespan studies
Analyzing 10,757 control mice across 18 years revealed unexpected year-to-year variation, altering the statistical significance of two rapamycin regimens when an outlier cohort was excluded.
18 years
Limit: Data came only from control mice in the Interventions Testing Program
Edited by Boris Djordjevic · AgingResearch.News
In a preprint analysis of historical data, researchers examined 10,757 control mice across 18 years of the National Institute on Aging Interventions Testing Program. The multi-site program tests potential lifespan-extending compounds across three locations using standardized protocols.
The authors evaluated lifespan variation within and across sites to build a predictive framework for identifying unusually long- or short-lived control groups. They found that year-to-year variation, especially in males, contributed substantially more to differences between cohort averages than expected from random sampling alone. When applied to the data, the framework flagged an unusually long-lived female cohort from 2022 and identified a previously unrecognized outlier male cohort, while finding that a suspected 2017 male cohort was within expected variation. Excluding the newly flagged male cohort changed the statistical significance of two rapamycin regimens.
Why it matters
Because lifespan studies rely on control animals to evaluate whether a compound slows aging, unrecognized shifts in baseline survival can alter whether an anti-aging intervention appears effective.
Caveats
The findings come from a preprint that has not yet been peer-reviewed. The work is a retrospective statistical modeling analysis of historical laboratory mouse data rather than a new laboratory experiment.
The paper
A prediction-interval framework for identifying outlier control cohorts in lifespan studies
University of Texas at San Antonio; The Jackson Laboratory
bioRxiv, Oct 22, 2026, Open access, CC BY, Preprint, not peer-reviewed
