[Personalised medicine requires a paradigm shift in the concept of evidence].

Personalised medicine via a central biobank will require the introduction of a paradigm shift navigating by detecting pattern and correlation, instead of evidence of limited use in our complex environment. Accordingly, to make sense, population-wide data comprising "deep phenotype" including genetics, must be submitted to evaluation by "machine intelligence", based on "unsupervised learning" - the algorithm concomitantly improving its power with increasing mass of data. A system not based on deductive logic cannot be checked by logic but must be taken on face value.