The geneticist whose claimed to find a new mechanism of genetic regulation is defending her work against critics.
Vivian Cheung of the University of Pennsylvania in Philadelphia says that her team stands by a May paper in which it reported that it had found more than ten thousand sites where transcribed RNA differed from an individual’s corresponding DNA sequence. The paper raised the possibility of an as-yet unknown mechanism that performs a new form of “RNA editing” in our cells. Cheung also says that new data from her own group and from others (PDF) supports the finding.
“We stand by our report that there are many sites in the human genome where RNA sequences differ from their corresponding DNA sequences, and the types of RNA-DNA differences (RDDs) are not restricted to the known A-to-G and C-to-U RNA editing events,” Cheung wrote in an email. “Since the publication of Li et al, we have analyzed more samples and included additional parameters in our analyses, the new data support and strengthen our initial findings.”
Cheung’s defense comes after Daniel R. Schrider, Jean-Francois Gout and Matthew W. Hahn at Indiana University in Bloomington published evidence earlier this month that 55% of the apparent mismatches reported by Cheung’s team are in fact accurate transcripts of the underlying genetic code, and thus do not represent RNA-DNA mismatches. Critics had pointed out this possibility and what they called other potential shortcomings of the work when Cheung’s original paper was published.
Hahn’s paper pointed out that more than half of Cheung’s team’s reported RNA-DNA differences were actually accurate transcription of paralogs – genes highly similar to, but not identical to – other genes. Some of these paralogs were present in the human reference sequence, and some were in the genomes of the individuals sequenced in Cheung’s original study. Cheung’s group did not find these paralogous sequences because they did not compare the transcribed RNA sequences to underlying DNA sequence from the people in the study or to the human reference sequence, Hahn said.
But Chueng writes that “the evidence presented by Schrider and colleagues are not rigorous.” She says, for example, that many of the paralogs identified by Hahn and colleagues in the reference sequence were not among her group’s reported RNA-DNA differences, and that Hahn’s group does not have evidence to back up its claim that as many as 90 percent of the originally reported RNA-DNA differences could be artifacts.
Hahn replies that his group found thousands of paralogs matching reported RNA-DNA “differences” in both the human reference sequence and in the genome sequences of the individuals whose DNA was examined in the study. And the hypothesis that 90 percent of the differences are incorrect is an extrapolation from known data, Hahn says.
Hahn also points out that another scientist, Heng Li, wrote on the blog Genomes Unzipped that he has compared the reported RNA-DNA differences from one individual in Cheung’s study to the individual’s fully sequenced genome. Li found that fewer than 50 RNA-DNA differences exist in that individual, far less than the 1244 that Cheung’s group reported. Most of these differences conform to previously known patterns of RNA editing, Li wrote.
Hahn points out that it was never in dispute that some editing of DNA transcripts occurs, but that Cheung appears to be softening her earlier claims that there were unknown editing mechanisms causing new types of RNA-DNA differences.
“[T]he language Dr. Cheung uses has shifted the argument from ‘there are 10,000 RDDs’ to ‘there are many sites that differ,’ ” Hahn wrote in an email. “Her paper did not get into Science because she found editing, but because they claimed it was so common and it used non-standard edits.” So although Cheung cites papers and abstracts that she says validate her work, Hahn says, “It is disingenuous to cite these papers in support of the Li et al. study in Science.”