I understand your premise, you would prefer that where no experiment has yet been performed, I shouldn’t propose a hypothesis. Especially if it runs the risk of sounding “holistic”. You interpret a claim where I clearly state possibility that needs experimental validation “Once again, data has not yet been produced on the direct effect of brazil nuts on human gut microbes…”.

To your credit, experiments need to controlled, so I could have avoided that critique by replacing the 2 studies that weren’t controlled with studies that were, even though the dialysis patient treatment was repeated in a controlled study.