Paired t−tests for ND50 means showed no significant difference between the Linear-bootstrap and Hill−MCMC methods ( Figure 2B , n = 50, t = 0.36, p = 0.72), whereas comparisons of Hill−MCMC vs Non−statistical (n = 50, t = 33.95, p < 0.0001) and Linear−bootstrap vs Non−statistical (n = 50, t = 33.95, p < 0.0001) were highly significant.
← all excerpts
CoreTIA: a modular, statistically robust transduction inhibition assay for AAV neutralization.
2
—
—
The sentences
When applying the Bayesian Practical Equivalence Test using θ = 0.3 log2 unit threshold ( Figure 2G, H ), both statistical methods produced ND50 estimates that exceeded the non-statistical estimates by more than this threshold, indicating practically significant differences under these conditions, with the Hill-MCMC estimate showing the closest alignment to the true ND50 value of 1/16 ( Figure 2G ).