In recent years, robustness of preclinical data has been increasingly in focus of discussions about the reasons why the success rate of new drug development is so low. There are various solutions suggested to enhance the robustness of preclinical data – stricter adherence to good research practice standards, collaborations and data sharing, etc.
It seems worth noting that the Big Brother, clinical research, has gone a long way to improve its quality standards. This illustrates what may also be the path forward for preclinical science. Two recent publication indicate important milestones that preclinical science may target as well:

The first one (link) appeared in The New England Journal of Medicine and discusses the final rule for trial reporting in clinicaltrials.gov. This well-known registry currently has more than 224,000 study records, of which only 23,000 display results information. The final rule presented by Deborah Zarin and her colleagues is aimed at enhancing compliance with the results-reporting requirements.
The idea of prospective trial registration is not unique to clinical science. In the preclinical world, there are also studies for which prospective protocol registration (along with the pre-specified endpoints) is highly desired. First of all, it is widely acknowledged that there is a major bias towards reporting of positive findings and one way to address this problem is to register important studies and require that, after completion, study results are reported as well. As Zarin et al. put it, „The time to decide whether a trial is worth doing is before the trial is started, not after participants have been put at risk“.

Another clinical science discussion that is worth highlighting refers to the use of statins (link). There is no doubt that statins are extremely used in prevention and treatment of cardiovascular disease. Evidence behind statins is deemed to be rock-solid and these agents have been used in tens of millions of patients worldwide for many years.
What does it take to raise questions about the use of statins today? And what kind of response one would expect to receive? A battle between the British Medical Journal and The Lancet has been triggered by a publication on statins’ side-effects. While the experts are discussing and will eventually come to a consensus agreement, it is interesting to note that, in preclinical sciences, such open discussion are not too common. It is often said that providing a direct and honest feedback on a peer’s work can be harmful because of the peer’s potential role of reviewer of your next grant application or a manuscript.  
Yet, one needs to acknowledge that self-correcting nature of science (link) requires such feedback. Thus, careful analysis of what makes clinical science different may be essential to identify the obstacles to create an open science culture.