Add mixed models to Milo differential abundance testing - #1069

Merged

Zethson merged 13 commits into

mainfrom

Aug 6, 2026

Commits on Aug 6, 2026

Add mixed models to Milo differential abundance testing
Zethson
committed
ceee591
Calibrate the mixed model p-values and speed up the fit
Zethson
committed
826858a
Reuse the random effect products and warm start the dispersion refit
Zethson
committed
482dcac
Correct the dispersion for the degrees of freedom the random effects spend
Zethson
committed
79f2917
Replace the results of a previous da_nhoods run instead of merging them
Zethson
committed
d823a91
Build the mixed model design with formulaic-contrasts and support contrasts
Zethson
committed
081026d
Subset the design frame once so a sample subset reaches every solver
Zethson
committed
4c8d028
Cover separated neighbourhoods reaching the results table
Zethson
committed
a51fae1
Estimate the dispersion by adjusted profile likelihood
Zethson
committed
4dc31c9
Let the effect of a variable vary between groups
Zethson
committed
48ae4c2
Point at da_nhoods when de_nhoods is given a random effect
Zethson
committed
46e893f
Drop random slopes from this pull request
Zethson
committed
bf70552
Bump tutorials submodule for the mixed model section
Zethson
committed
7b9f794