In this study, the authors use bioinformatics tools to propose potential new drug therapies for autism. They separately analyse the alterations in gene expression (or gene products) that occur most commonly in autism and compare them with those produced by a selection of drugs already approved for other conditions and disorders. In this way, they select four drugs that could compensate or reverse the alterations produced by autism.

However, the authors seem to overlook or at least underestimate the tremendous complexity of this neurodevelopmental disorder, in which multiple factors are involved, not only genetic but also environmental, since they reduce the complex molecular alterations involved in autism to a small network of interactions between proteins (or gene products), which they use as a basis to identify the candidate drugs.

On the other hand, among the four drugs finally selected, they include: an opioid (Loperamide), which since it does not reach the brain could hardly produce the supposed therapeutic effects, but still could cause peripheral adverse effects and therefore is not recommended for children (especially under 6 years old); a female sexual hormone (progesterone) and a contraceptive (drospirenone), about which it is not difficult to understand why they are not recommended as treatment in children who have not even initiated the sexual development. It is therefore quite unreasonable to propose these drugs for the treatment of social deficits in children with ASD without first testing whether they are safe and effective in animal models.

Although bioinformatics is proving to be a relevant complementary method to animal research in many fields, these studies show that it is still far from being an alternative.