Follow-up from #25:
The performance issue is pre-existing due to the fitting logic in the stats section. In looking into this I noticed there are other issues where the current implementation is different from our reference examples.
- Grouping by chemical and sex for fitting?
- Including controls at 0 concentration?
- Limit which curves are shown based on significance?
Follow-up from #25: