Are You Sure?
31/07/12 12:55 Filed in: GenomeWeb Daily Scan
Submitted by S. Pelech - Kinexus on Tue, 07/31/2012 - 14:59
The increased rate of retraction of scientific research results in publications actually reflects a wide variety of confounding factors. Scientific manuscripts may be submitted prematurely for publication due to the limited resources available to the authors and the high pressure to demonstrate productivity for their continued funding. The effect is magnified when major institutions are quick to widely publicize the scientific results of their researchers in public relations campaigns to justify their continued support. It is suspected that there is a much higher degree of plagiarized and even fraudulent data from some countries more than others, however this clearly also happens in North America and Europe. It is also probably much easier to detect copied data with today's increasingly powerful Internet search engines than before.
As biomedical research has becomes increasingly diversified and specialized, it becomes more difficult to obtain highly informed and rigorous peer-review. But with the proliferation of new scientific journals, it is even easier to publish work that has been previously rejected. Those involved in systems biology research can fully appreciate the amazing amount of variation between individuals in populations and even within their own cells with respect to gene sequence, expression and protein post-translational regulation. This means that reports of specific biomarkers for biological processes are becoming increasingly challenged as more data accumulates to the contrary.
The various tools and procedures that are used to generate the research data also have severe limitations. For example, the vast majority of commercial antibody probes sold for biomedical research are non-specific and impotent. Microarrays with oligonucleotides, antibodies, proteins and peptides can generate high rates of false positives and false negatives. With higher rates of genome-wide sequencing comes increased sequencing errors. Mass spectrometry likewise has a whole set of artefact and accuracy issues. As we push the limits of these powerful tools, we have to expect that questionable results will be generated. It is critical that key findings are reproduced with complementary or alternative methods and larger numbers of samples.
The one saving grace about scientific research is that it is self-correcting process and the truth still ultimately emerges.
Link to the original blog post
The increased rate of retraction of scientific research results in publications actually reflects a wide variety of confounding factors. Scientific manuscripts may be submitted prematurely for publication due to the limited resources available to the authors and the high pressure to demonstrate productivity for their continued funding. The effect is magnified when major institutions are quick to widely publicize the scientific results of their researchers in public relations campaigns to justify their continued support. It is suspected that there is a much higher degree of plagiarized and even fraudulent data from some countries more than others, however this clearly also happens in North America and Europe. It is also probably much easier to detect copied data with today's increasingly powerful Internet search engines than before.
As biomedical research has becomes increasingly diversified and specialized, it becomes more difficult to obtain highly informed and rigorous peer-review. But with the proliferation of new scientific journals, it is even easier to publish work that has been previously rejected. Those involved in systems biology research can fully appreciate the amazing amount of variation between individuals in populations and even within their own cells with respect to gene sequence, expression and protein post-translational regulation. This means that reports of specific biomarkers for biological processes are becoming increasingly challenged as more data accumulates to the contrary.
The various tools and procedures that are used to generate the research data also have severe limitations. For example, the vast majority of commercial antibody probes sold for biomedical research are non-specific and impotent. Microarrays with oligonucleotides, antibodies, proteins and peptides can generate high rates of false positives and false negatives. With higher rates of genome-wide sequencing comes increased sequencing errors. Mass spectrometry likewise has a whole set of artefact and accuracy issues. As we push the limits of these powerful tools, we have to expect that questionable results will be generated. It is critical that key findings are reproduced with complementary or alternative methods and larger numbers of samples.
The one saving grace about scientific research is that it is self-correcting process and the truth still ultimately emerges.
Link to the original blog post