Stats cutoffs are anyway arbitrarily set. Just because there’s a significance with a conservative test doesn’t mean there’s a biologically relevant effect (or size) and vice versa. Stats are a tool for discovery and if they have validated their findings with functional experiments I see no issue.