This research focuses on a special case of General Purpose Technology: Bioinformatics. It explores whether – and to what extent – Bioinformatics inventions build upon inherently diverse knowledge sources. Precisely, the role of scientific and technological diversity (measured with Shannon-Wiener diversity index) as driver of impactful Bioinformatics inventions (measured at different standard deviations of the forward citations distribution) is investigated. To this purpose, we carried out an analysis of both Non-Patent and Patent references cited into Bioinformatics patented inventions in the period 1976-2014. Results from a series of logistic regression models indicate that different degrees of impact require different degrees of knowledge diversity; at the same time, and importantly for practitioners and scholars, recombining diverse scientific and technological knowledge bases not always lead to impactful inventions. In other terms: the interplay of science and technology is not always the best option to get impactful inventions.