What should MetaManifold do next? Bring your wish list

[hyperpolymath]

This is the front door for new ideas — the small convenience that would save you an hour, or the whole analysis method the tool is missing. Both are wanted.

What helps an idea travel here:

• the job it does for you, not the mechanism you imagine — "I want to know which taxa differ between my two groups, with a defensible multiple-comparison story" travels better than "add DESeq2";
• what you currently do instead, and what it costs you;
• whether it belongs in the pipeline (denoising onward), the analysis layer (diversity, ordination, differential abundance), or the UI.

Ideas that gather interest become tracked issues with the analysis-workflow spelled out, and the good ones get built. Nothing here is a promise — but everything here gets read.

Existing direction, if you want to know where offers would land: differential abundance currently offers a negative-binomial GLM and a CLR-based Welch test; ordination has NMDS with PERMANOVA over blocks; the roadmap's open questions lean toward constrained ordinations, occupancy-style models for imperfect detection, and multinomial/Dirichlet-multinomial regression. Arguments that those priorities are wrong are exactly as welcome as new ideas.