Scientists seek demigod status, journals want blockbuster results, and retractions are on the rise: is science broken?
On 5 August 2014, a celebrated Japanese scientist was found dead, hanging by his neck at his workplace, his shoes politely removed and placed on the landing of the stairs. Yoshiki Sasai, 52, was a legendary stem-cell expert, widely regarded as an exceptional scientist, who worked at the RIKEN Center for Developmental Biology in Kobe. Seven months before he killed himself, Sasai and colleagues in Japan and Boston announced a stupefying research breakthrough in two papers in Nature. They claimed that ordinary mouse blood cells could be transformed into powerful stem cells – the holy grail of regenerative medicine – by simply bathing them in a mildly acidic solution (called STAP, for stimulus-triggered acquisition of pluripotency).
Almost instantly, the work was called into question. Accusations surfaced in the science blogosphere that images in the papers had been duplicated or altered, and at least eight scientists announced that they were unable to reproduce the experiment. In February 2014, RIKEN launched an internal investigation, and found the 30-year-old lead author, Haruko Obokata, guilty of scientific misconduct (which includes falsification, fabrication, or plagiarism). She had been Sasai’s protégé.
In June, Science reported that earlier versions of the STAP work had been rejected by three top journals: Cell, Science and even Nature itself. Science quoted RIKEN’s report, where peer reviewers raised many troubling questions. ‘This is such an extraordinary claim that a very high level of proof is required,’ wrote one. Another said the paper was ‘simply not credible’. Scientists once again took to the blogosphere asking why Nature had published flawed work and whether journals today value hype over substantive science.
In July, Nature retracted both papers – essentially stamping them with a scarlet letter. Retraction lofted the scandal to worldwide infamy. One evening, when Obokata left work in a taxi, a reporter on a motorcycle started following her. She stopped at a hotel, but was pursued up the escalator and into the bathroom by five journalists, including a cameraman, and sprained her right elbow trying to get away.
Then, in August, Sasai killed himself. This was in spite of the fact that he’d been cleared of fraud, and his share of responsibility was linked to lack of proper oversight. In a suicide note to his family, he wrote that he was ‘worn out by the unjust bashing in the mass media’. His brokenhearted family simply said: ‘We feel crushed by a deep sorrow… we see nothing but despair.’
That same month, RIKEN also announced it was conducting a second investigation into possible laboratory contamination during the experiments. The institution’s reputation was still so damaged that in late October six top administrators volunteered to atone by returning between one and three months of their salary.
Based on this narrative, one might conclude that retraction is a near-perfect guillotine, heartless perhaps, but an alarmingly potent tool for self-correction in science.
That is not the case. The STAP story is a tale of all that’s troubled in the scientific enterprise today: scientists seeking demigod status and flying too close to the sun with their claims; journals smitten with a potential blockbuster finding, and overlooking vexing questions ahead of publication; retractions on the rise, entering mainstream awareness, and leaving an entire scientific community frightened of the resulting stigma.
Retraction was meant to be a corrective for any mistakes or occasional misconduct in science but it has, at times, taken on a superhero persona instead. Like Superman, retraction can be too powerful, wiping out whole careers with a single blow. Yet it is also like Clark Kent, so mild it can be ignored while fraudsters continue publishing and receiving grants. The process is so wrought that just 5 per cent of scientific misconduct ever results in retraction, leaving an abundance of error in play to obfuscate the facts.
Scientists are increasingly aware of the amount of bad science out there – the word ‘reproducibility’ has become a kind of rallying cry for those who would reform science today. How can we ensure that studies are sound and can be reproduced by other scientists in separate labs?