The vulnerability of science has been highlighted by the unfolding controversy over a paper published in Nature on transgenic introgression into Mexican maize1. The scientific enterprise is labor intensive, and unforgiving of carelessness. In practicing science, we must contend with the difficulty of abandoning a favorite hypothesis when data do not support it. We also have to shield objective investigation from political, financial, and ideological diversion or risk trying to prove something that is less than true. External agendas can also detract from science, with manifestations ranging from manipulation of scientific debate for political ends, to wanton destructive actions. Concerns about the impingement of financial influences on research have presumably motivated recent editorial policy revisions by journals such as Nature and Nature Biotechnology requiring authors to disclose competing financial interests.

To overcome these challenges, we require both the support and scrutiny of our colleagues. For scrutiny, peer review is often heralded as the gold standard. Non-experts in a field often rely on the filter provided by peer review to screen for quality and originality of research. To ensure the integrity of this process, reviewers must be expert in the experiments from which data are drawn and not be predisposed to concur with (or disfavor) the conclusions in question. Reviewers must be expert and unbiased enough to rigorously assess not only whether authors made relevant use of scientific reagents, but also whether they have dutifully employed the scientific method, produced information of importance to a field of inquiry, and adequately addressed current knowledge in the field. Many influences can press for a lowering (or raising beyond reach) of the bar in the stringency of peer review for particular studies, including socioeconomic concerns, the lure of prominence in the popular media, and political agendas. But to compromise this standard weakens the fabric of science as a self-correcting, knowledge-building process.

However deeply we depend on peer review, it is not an end point, but rather the debut of hypotheses before a body of knowledge. Even properly conducted peer review can err. Thus, the re-examination of published data and the review of newly generated data is essential to the health of science. Personal clashes are tangential; the evidence (or lack thereof) will tell the story.

The study published in Nature reporting transgenic DNA found in traditional Mexican maize1 has drawn into the spotlight the adequacy of current peer-review policies and the role of a critical scientific audience. Questions have been raised about the relevance of the study's conclusions and the validity of the data from which they were drawn2. Critiques of the study3,4 have prompted the editors to conclude that the original study was not fit for publication. These activities emerge as part of proper scientific discourse. There have also been missteps, such as the release of a reviewer's comments urging retraction of the study5 before the journal's editorial embargo was lifted.

To those removed from the rigors of scientific scrutiny, this heated response might seem intimidating. In fact, the “Joint Statement on the Mexican GM Maize Scandal” released by Food First6 (Oakland, CA) interpreted the criticisms as “intimidatory tactics” and stated, “Pro-industry academics are engaging in a highly unethical mud-slinging campaign.” Regardless of less-than-civil statements from both sides, scrutinizing data and conclusions stands as a duty for scientists.

The “Joint Statement” goes further to demand that the scientific community “censor those academics and institutions that slander the competence or integrity of those who publish peer-reviewed studies.” As noted before, peer review is the beginning of a critical process, not the end, and criticism is not slander. Critical examination from our peers, whether it corrects or validates our conclusions, will bring us closer to fact. Criticism is welcomed when pursuit of knowledge is the guiding priority. Censoring reasoned scientific disputes would invite a culture of mediocrity in which suspect conclusions would go politely unchallenged. A worst-case example of this can be found in Stalin's appointee Trofim D. Lysenko, who devastated Soviet agriculture and biological science with such policies.

In principle, peer review overrides political, ideological, and financial influences, and other non-scientific considerations. Functional peer review evaluates scientific data and conclusions on scientific criteria alone. Therefore, it might be argued, editorial policies requiring the disclosure of potential financial conflicts of interest are superfluous: if such disclosure is necessary, then it follows that the peer review process must be faltering.

Editors and reviewers are, in effect, censors of scientific disclosure. They are no less susceptible to conflicts of interest than are authors. Is it then the secrecy of the editorial process and the anonymity of peer review that require baring to the public eye? Readers may gain a good deal more insight from the transparency of the censors than from knowledge of the minutiae of authors' commercial associations.

Whether or not we doubt the validity of a colleague's conclusions, we cannot ignore the importance of fellow scientists' pursuit of knowledge. Our esteem for knowledge should motivate us to give both our criticism of data and our support of efforts to generate data. It is unconscionable for a member of the scientific community to give complicit or implicit backing to suppression of criticism or destruction of research.

As scientists, we must be relentless in critically examining data, especially our own. We must show unyielding opposition to destructive acts, both physical and political, against research. And we should not be complacent about deficiencies in the systems that mediate scientific discourse.