## Problem 291 — rank:48

**Title:** Metabolism-first versus genetics-first

**Added binding assumptions (LaTeX):**

```latex
Model classes can overlap, so experimental discrimination must be quantified on separated observable subfamilies or acknowledge a shared equivalence region. Fix within-class priors and complexity before calculating marginal likelihoods. A laboratory result can establish chemical feasibility or reject a proposed pathway without determining which pathway actually occurred historically. The endpoint, feed restrictions and observation model must be common to both classes.
```

**Direct edits:** 0. **Fresh external checks:** None; supplied sources retained as provenance.