All the time, I have believed that MR analysis can provide the information such as "there is a causal effect of X on Y", until I saw a statement on this paper (https://www.nature.com/articles/s41588-020-0631-4):
A low P value from CAUSE (or any MR method) should not be regarded as proof of a causal effect. Instead, it is an indicator that the summary statistics for the two traits are consistent with a causal effect.
I am not quite sure what does this statement mean. Does that mean even I have a p-value which is lower than 0.05, but I still cannot say "there is a causal effect of X on Y"? If that is the case, then how should I interpret the result with low p-value?