PerspectivesAre you interested in submitting a Perspective Article? Be sure to read The Science Advisory Board's Editorial Guides for Perspective Articles. Click here. What's in an image? by Wim D’Haeze, Ph.D. As recently as a decade ago, scientists relied on relatively old-fashioned methods to present their scientific data. For example, images observed using a microscope were captured by a conventional camera. The pictures needed to be developed, and the authors had to make collages that then needed to be photographed again prior to the submission of glossy prints to a journal by mail. Preparing a poster was even more cumbersome (and expensive) as the enlarged glossy prints needed to be glued on large poster paperboards. Fortunately, things have changed and most of the data collection and presentation has moved into digital format. Professional software packages are available that allow the scientist to store images in compressed formats, to drag those images into other text processing or presentation programs, and to efficiently combine them in a single document that contains the labels and indications that meet the journals’ guidelines. Obviously, these new developments provided the scientific community with many advantages. It is now possible to generate numerous digital microscopy images per day, all taken under preset parameters; this makes it possible to quickly and efficiently obtain the data required to answer a given biological question. This electronic flow of data also affects the submission of manuscripts to journals. Most journals have implemented web-based submission systems that substantially facilitate the transfer of manuscripts to peer-reviewers around the globe. The major advantage of the imaging software packages such as Photoshop, in my opinion, is that certain settings of difficult-to-take images can be adjusted (e.g., contrast, intensity of parts of the image, etc.). For example, dark-field-optics microscopy images often contain dust particles that appear as shiny spots on the image. Photoshop software allows for the removal of those spots, thereby improving the quality of the image without changing its original meaning. Unfortunately, some recent reports of scientific deception have prompted the scientific community to question whether a published image does represent the actual data. Indeed, the downside of modern image technology is that authors are now in the position to fabricate or alter their control images, to remove (or add) bands to a protein gel, to duplicate lanes or combine lanes of different protein gels into one image, and to remove or adjust backgrounds. Several cases of scientific fraud have recently caught the public’s attention and the impact is greater than one may expect. Not only will the reputation of the journals (often of high impact) that published the counterfeit manuscripts be affected, the entire institution affiliated with the publication will be affected. Deceptive scientists could face criminal charges and possibly, depending on the country, a period in jail. In addition, researchers will attempt to reproduce the faked data or to perform similar experiments applicable to their research, which could negatively impact the careers of numerous scientists. In such cases of fraud, the first question in my mind is “Why?” Why would so-called “top” scientists be part of such deceitful practices? It is possible that once in a while an innately dishonest individual becomes a scientist; however, the path to a doctoral title is usually stringent, with many checks and balances, making this scenario rather unlikely. What seems more likely to me is that today’s extremely competitive environment -- along with substantial reductions in the funds allocated to scientific research -- puts some principal scientists in a difficult and stressful position. They may think that data forgery could hasten publication in a high-impact journal, which would temporarily solve some of their concerns. Obviously, blatant alteration of scientific data is a criminal practice and one of the major responsibilities of journal editors and peer-reviewers is to catch those cases. But how? When researchers seek to remove an object from an image, they often camouflage it with a patch of nearby background that is not so easy to discern by eye, but can easily be detected using mathematical methods. Pioneered by the Rockefeller University Press, the Journal of Cell Biology has set up standards for acceptable manipulation of images. Authors are urged to submit original images in order to allow the journal to investigate whether these standards have been violated. By utilizing this procedure, several cases of deceitful misinterpretation or misconduct have been caught. Although it is not easy to develop a procedure that will identify all such cases, this is a promising step in the right direction. This procedure would have caught some of the most recently reported cases of scientific fraud prior to publication. The drawback of this method is that it will only determine unacceptable changes to an image after the image has been taken. In other words, authors are still able to extensively adjust microscope settings to mask or enhance particular observations, or to provide unrepresentative images that show what the authors would like to be shown. Such deceptive actions will likely not be detected by this method. In addition, many life science manuscripts do not contain microscopy-related images and rely on data presented in the form of graphs and/or tables. In these cases, a close look at the data populations and respective statistical analyses may also provide information regarding the correctness and scientific quality of the data. However, it is still difficult to detect possible fraud. Numerous instances of the publication of falsified scientific data strongly suggest that the peer-review and editorial process commonly used by top-tier journals needs to be improved considerably in order to prevent the publication of fraudulent data. It is my opinion that the peer-review process in its current form will continue to allow certain fraudulent manuscripts to fall through the cracks. Most often, the journal’s editorial board makes a first selection among the incoming manuscripts and only those manuscripts that represent a major novel contribution of interest to the broad scientific community proceed to the next step. The next step is a critical evaluation of the quality of the data and conclusions by premier peer-reviewers who are often senior professors with extremely busy schedules. Even in cases where the submitted images are being compared with the original ones, it remains difficult to detect misconduct if the peer-reviewers and editors continue to rely on what’s included in a submitted manuscript and possibly some “raw” data without having thoughtful discussions with the authors. If some of the data presented in a manuscript seems to be questionable as determined by the editor, members of the editorial board, and the external peer-reviewers, it might be very helpful to have a conference call with the major authors to thoroughly discuss the manuscript. It is relatively easy to truncate data and conclusions in written form, but it is more difficult to support falsified data while orally discussing the subject with a committee of specialists. Of course, such enhanced peer-review process would require considerably more time from the editor, members of the editorial board, and the external peer-reviewers and might also require additional costs. However, it has the potential to save considerable time and funds that could be used to support honest researchers instead of wasted by sham scientists. Wim D’Haeze is a Bio-Engineer in Chemistry and received his Ph.D. in Biotechnology at Ghent University (Belgium) in June 2001. His doctoral thesis work was focused on the understanding of several early steps of the symbiotic interaction between the Gram-negative soil bacterium Azorhizobium caulinodans and the tropical legume Sesbania rostrata. These initial steps require the production of bacterial compounds including signal molecules and complex surface polysaccharides that are pivotal for invasion of the plant tissue and the formation of new organ tissues. In the three subsequent years, he performed post-doctoral research at the Complex Carbohydrate Research Center at the University of Georgia (Athens, GA) dealing in part with the structural and functional characterization of azorhizobial extracellular polysaccharides. Currently, Dr. D’Haeze is employed at The Scripps Research Institute (La Jolla, CA) as Science Writer and focuses on a new horizon regarding the molecular basis of devastating neurodegenerative diseases, such as Alzheimer’s and Parkinson’s diseases, in order to screen for and develop new therapeutics. E-mail: wim.dhaeze@sbcglobal.net ### << Previous Next >> [ View All Perspectives ] |
|
