Tangential Thoughts: Robot Scientists for a Better Use of Existing Data – And Why Translational Science May Still Need a Slightly Different Approach

Is serendipity necessary for innovation? Or in other words: Would an autonomous scientific discovery process that utilizes all available data at the time be incapable of innovation? Some think so. But not researcher Andrew Sparkes and colleagues who created Adam and Eve, two robot scientists, designed to carry out biomedical scientific research. The researchers claim that scientists robots will “make scientific information more accurate, reproducible and reusable”.

Adam and Eve are capable of generating hypotheses about a problem based on information obtained from publicly available databases, designing experiments to test these hypotheses, running the physical experiments, analyzing, interpreting the resulting data – and they even collaborate. Eve, for example, is a prototype system to demonstrate the automation of closed-loop learning in drug-screening and design.

So why not stretching this idea a bit? Could such a robot help support the clinical and translational research process? The authors of the recent paper “Translational Medicine – doing it backwards” may disagree. They argue that the general approach to hypothesis-driven research poorly suits the needs of translational biomedical research “unless efforts are spent in identifying clinically relevant hypotheses”. As Steinman pointed out, animal models, for example, can lead to results that are the opposite of what is ultimately seen in human disease. So, the authors propose “that hypothesis tested research should follow ‘factsdriven research’ and only when the collection of facts relevant to human disease has been extensive, should hypotheses be constructed to expand beyond what can be directly observed. What is needed is an approach that begins at the Bedside and then goes to the ‘Clinical Bench’.”

I guess once there are public databases available filled with “clinical realities” provided by clinically active physicians and non-physicians, robots like Adam and Eve could frame their research questions accordingly and reverse the discovery process starting with the “human reality”.