Scientists and journal editors are still struggling to resolve the confusing saga of the reactome array, a chiplike device intended to easily analyze all the enzymatic activity inside cells. First unveiled last year in a paper in Science, the reactome array was quickly declared "impossible" by biochemists who found its description and supporting data full of flaws. The uproar led Science to publish an "Editorial Expression of Concern," and the authors' institutions launched an investigation.
Now, as originally reported by Nature last week, an ethics committee from the Spanish National Research Council (CSIC) has recommended that the paper be retracted, concluding that the "content of the publication does not have all the necessary experimental support for the conclusions reached." But one of the authors of the paper maintains that the reactome array works. And several scientists who have used the array, including a Nobel Prize-winning biologist who has subjected it to "blind" testing, also defend the device and suggest that its developers are guilty of little more than poor manuscript preparation and communication.
"This technique is a major breakthrough in high-throughput methodologies that will become useful for researchers in many scientific fields," says Antonio Suárez García of the Center of Biomedical Investigation in Armilla, Spain, who says he has used reactome arrays to characterize differences among lean and obese people in gut enzymatic activity and microbial populations.
The reactome array—as described last year in the 9 October issue of Science by Manuel Ferrer of the Spanish National Research Council's (CSIC's) Institute of Catalysis in Madrid and colleagues in Spain, Germany, and the United Kingdom—consists of a glass slide peppered with 1676 synthesized compounds representing cellular metabolites and other targets of enzymes. Each one, the authors said, is linked to a dye that emits light when the compound reacts with an enzyme. The reactome array in theory would allow a quick analysis of all the enzymes present in a bacterium, cell, tissue, or other biological sample.
Immediately after the Science paper was published, researchers objected that the description of how the fluorescent dye attaches to each compound was flawed. And biochemists almost as rapidly picked holes in other parts of the paper, complaining especially about the lack of information and mistakes in the supplementary online material meant to detail how Ferrer and his colleagues synthesized the many hundreds of enzyme targets. Science didn't escape the fierce criticism, as some condemned its peer-review process, especially after the journal acknowledged that the paper's primary reviewers did not include a synthetic organic chemist.
The CSIC Ethics Committee (CEC), led by maize geneticist Pere Puigdomènech of CSIC's Institute of Molecular Biology in Barcelona, has now delivered its own verdict in a brief letter sent to Science dated 13 July. The panel concludes that:
Clear indications of deviation of good scientific practices in the lack of controls in the experiments, in the treatment of data, in the lack of information in the publication and in contradictions in the responses to the journal appear to have been produced by the authors, including researchers belonging to CSIC.
The committee doesn't provide any further details to back up its retraction recommendation and doesn't explicitly conclude that the reactome array does not work; the letter in fact notes that "a number of scientists are convinced of the validity of the methodology described."
That mixed message may be attributed to the fact that the CEC investigation was relatively limited in scope. "In Spain, CSIC researchers work under the rules of Public Administration, and to investigate individual responsibilities a formal investigation that may include a search in the laboratories and research books and might lead to disciplinary actions has to be decided by the president of CSIC. He is waiting for the reaction of Science in order to take his decision. This kind of action is outside the mandate of CEC," Puigdomènech tells ScienceInsider.
The ethics committee's report has disappointed some scientists. "I frankly think the CSIC Committee could have done a much better job by clearly distinguishing between the part of the work that is entirely sound from the bits and pieces which could be doubtful or bear anomalies. To this end, they should have interviewed and site-visit[ed] the Authors at stake, which they did not do -the whole process was basically done at a distance," Victor de Lorenzo of CSIC's National Center of Biotechnology in Madrid e-mailed ScienceInsider. He continues: "If the paper is just withdrawn or retracted, I am afraid of two consequences [i] the wider scientific community will be deprived of the use of an incredibly powerful tool, further amendments and improvements notwithstanding and [ii] any smart cookie (maybe some of those who raised the most venomous criticism) may now rush, if they have not done it yet, to replicate the work, produce a publication and claim credit for themselves."
Like García, De Lorenzo considers himself a satisfied user of the reactome array. "We have employed the chips before and after the controversy and—to the best of our understanding—they generally work as anticipated. The instances that do not fit the expectations might be indicative of new, unprecedented metabolic reactions, which of course need to be confirmed separately," he says.
The biggest defender of the reactome array may be Nobel laureate Richard Roberts of New England Biolabs in Ipswich, Massachusetts. Initially skeptical of the technology, Roberts visited Ferrer's lab and came away impressed enough that he says he has urged Science to avoid rushing to any judgment. Roberts has sent Ferrer's lab anonymous samples of 10 purified proteins, including metabolic and DNA-modifying enzymes. The Spanish lab analyzed each solution with the reactome array and sent the data back to Roberts. When he unblinded the results, the "appropriate" substrates had fluoresced on the reactome arrays in eight of nine cases. (Technical problems stymied analysis of one solution.) "Based on our test, I am convinced that the reactomes work. We have actually engaged in collaboration with Ferrer to study H. pylori," says Roberts, who has submitted his data to the editors at Science evaluating the reactome array paper.
While acknowledging that the Science paper and its supplementary online material had multiple errors for which he takes responsibility, Ferrer does not agree that the publication should be retracted. If it is, however, he plans to continue his efforts to win over skeptics. "I will do my best to convince the scientific community that the chemistry and content of the paper are correct through additional blind experiments and future articles, subjected to peer review by organic chemists, describing in detail the chemistry," he e-mailed ScienceInsider.
Ferrer, Roberts, Puigdomènech, and many others are waiting to see how Science responds to the CSIC recommendation. A journal spokesperson last week released the following statement: "Science has received the committee report and is now following up with a few additional questions, including some to the other institutions involved and [we] hope to have a published statement and decision as soon as possible."