Couzin and Unger (1) discuss how suspect data are continued to be cited several years after publication. I would like to present an example of ongoing propagation of apparently misinterpreted data, published in this journal (2), on identification of a novel ammonium transporter from soybean.
In a Letter to Science submitted in September 1998, we drew attention a to striking similarity of this putative transporter to transcription factors, provided an alternative interpretation, and suggested how this could be verified experimentally. The expert reviewing both our manuscript and the authors’ rebuttal acknowledged our concern but also recommended against publication of a comment questioning validity “solely based on sequence gazing.” Science declined to publish our comment arguing that “interpretations and insights, then possibly new experiments, will happen after a Report is published in Science, and usually, this work will appear in specialty journals.”
Admittedly, this is what happened but it turned out that self-correction by the community was slow and not penetrating all affected research areas. Our concerns about inappropriate data interpretation were soon confirmed by an in silico analysis similar to ours and experimental evidence based on the yeast mutant complementation proposed in our comment, published in a leading microbiology journal (3). We discussed the conflicting data in a review on ammonium transport, published in a plant physiology journal (4). The regulator hypothesis was briefly acknowledged in a review (5) co-authored by one of the Science paper authors, and found support by an independent phylogenetic study published in an evolutionary biology journal (6). However, of the 30 other cites in the period after publication of the conflicting experimental evidence (3), only three papers and one self-citation refer to this. In the other Review and research papers, the protein is still erroneously described as an ammonium transporter. This misleading paper trail, which also runs through other types of publications like PhD theses, symposium proceedings, and web pages, is likely to endure for some time.
The collateral damage caused in this particular case may have been minimized if the issue had been raised timely in Science, if not through our whistle-blowing comment and authors’ rebuttal, certainly following independent experimental revocation.
References
1. J. Couzin, K. Unger, Science 312, 38 (2006).
2. B. N. Kaiser et al., Science 281, 1202 (1998).
3. A.-M. Marini et al., Mol. Microbiol. 35, 378 (2000).
4. A. Van Dommelen et al., Aust. J. Plant Physiol. 28, 959 (2001).
5. D. A. Day et al., Cell. Mol. Life Sci. 58, 61 (2001).
6. M. J. Buck, W. R. Atchley, J. Mol. Evol. 56, 742 (2003).
E-Letters: ![[Read E-Letter]](/icons/misc/sectionDOWN.gif){kind=icon}
