In last week's This Week In Science, I summed up a paper published in Scicence by Li et al.1 which was hyped up as being a challenge to the Central Dogma of molecular biology. The authors compared the RNA sequences from numerous individuals to the original DNA sequences they were derived from. What they found were a multitude of sites where the RNA sequences differed from that which would be expected given the original DNA sequence. Moreover, these variations were shared across individuals, indicating that they were not likely due to random mutation. Furthermore, the team found proteins that matched the varied RNA sequences and not the DNA sequences. These results suggested that there exists some yet unknown editing step during transcription that alters individual nucleotides in the resulting RNA transcripts.
However, declaring the Central Dogma to be toppled may have been a bit premature. According to Lior Pachter at the University of California, Berkley, the variations discovered by the researchers could be artifacts caused by their sequencing equipment. Many of the sites that were found to contain altered nucleotides lie in regions which are known to often cause RNA sequencing errors. In other words, the variations that the team observed, in many cases, might just be sequencing mistakes.
Further skepticism has been shown by Joe Pickrell at the blog Genomes Unzipped (which I highly suggest reading, as he goes into quite a bit more detail than what I've presented below). He points out that the differences in RNA and DNA that the authors discovered might be false positives created by attributing a particular RNA sequence to the incorrect DNA sequence. For any given DNA sequence, there are bound to be other sequences very similar - even almost identical - to it2. If you are given an RNA sequence, then, how do you determine which of the very similar DNA sequences it is derived from? Unless one takes steps to remove the incorrect sequences, it is very likely that you will end up with a false-positive. It would appear that Li et al. did not take such steps.
Pickrell points out another problem with sequencing and mapping through RNA splice sites. Mammalian genes are frequently alternatively spliced, and a cDNA library like the ones Li et al. used will have multiple isoforms of a gene. When mapping such transcripts back to the genome, you have to keep in mind that the genomic sequence will still contain the introns that have been excised in the mature mRNA transcripts. If you compare the shorter, edited mRNA to the longer, unedited DNA, you're likely to find many differences between the two. Mapping a particular sequence read to the wrong isoform will generate false-positives. Pickrell shows that Li et al. did just this on at least one occasion.
So widespread RNA editing in humans might not be a reality. It's possible that it is, but problems with the procedure used by Li et al. raise many doubts. I'm looking forward to reading the follow-up research. Until then, as perhaps Mark Twain would say, reports of the death of the Central Dogma have been exaggerated.
1. M. Li et al. "Widespread RNA and DNA sequence differences in the human transcriptome". 2011. Science. doi:10.1126/science.1207018
2. It should be obvious that the larger the given sequence, the fewer (near)identical sites there will be.