I have always found it problematic, when reading articles with far reaching conclusions on the basis of solely computation-based GO analysis which are not validaded and are based often on GOs of organisms phylogenetically far removed from the studied non-model species, to attach meaning to the conclusions. In view of this, it was interesting to read the paper by Friedrich et al. (2019) Bioessays 41, 4 article no. 1800169 on microRNA targets of especially non-model organisms, where they say for example "GO enrichment analyses, which are often performed on experimentally nonvalidated targets, are most likely meaningless". Why is it so common to do just this and thereby obscure getting real biological information? Is it just because it is relatively easy, and actually validating the target and the functional significance of the response is difficult?