Tuesday, March 31, 2009

Way too Sensitive?

The techniques for detecting DNA and RNA are extremely sensitive. This sensitivity often leads to misinterpretations because it become difficult to separate signal from noise. The idea that 90% of our genome may be transcribed into functional RNA, for example, may be due to the sensitivity of an assay that can easily detect tiny amounts of accidental transcription.

Similarly, the often proclaimed ubiquity of alternative splicing may be due to the easy detection of splicing mistakes. Other examples of problems with noise might be the presumed abundance of small regulatory RNAs and the frequency of transcription factor binding sites.

The problem is acute when it comes to analyzing DNA from fossils. There, tiny amounts of contamination can really screw things up. That's why John Hawks is also interested in this problem of over-sensitive DNA assays.

Just how sensitive is the technology? Hawks has found a very interesting and informative example [The trouble with contamination]. This example is about contamination but keep in mind that it also applies to the detection of noise in transcription, DNA binding, and splicing.

Here's the original report from the BBC ['DNA bungle' haunts German police]. Over the past few years German police have been on the lookout for a mysterious woman who was linked to several murders. Her DNA was found at over 40 different crime scenes. This woman became one of the most wanted people in Europe but nothing was known about her aside from her DNA.

Finally someone became suspicious and started to look closely at the way they were collecting and analyzing DNA. To make a long story short, the alleged murderer is a factory worker in Bavaria who works in a factory that manufactures cotton swabs. The same swabs that are used to collect samples at a crime scene. Those swabs were contaminated with her DNA.

There's a lesson here. Any technology that can detect the DNA from a factory worker on a cotton swap is quite capable of detecting tiny insignificant amounts of nucleic acids inside a cell.


  1. Regarding the ubiquity of alternate splicing, i have asked a molecular biologists/biochemist principal investigator on whether it is a true phenomenon. I have posed your doubts on alternate splicings in conserved proteins in metabolic pathways (such as kreb cycle), but his reply is that from what his lab has found is that there is generally alternate splicing forms. The point is that I don't know who to believe....

  2. anonymous says,

    The point is that I don't know who to believe....

    Look at the data yourself, if possible. Ask your friend to give you the list of variant proteins for each of the genes of the citric acid cycle. Make up your own mind about whether those variants make sense or not.

    For each enzyme of the citric acid cycle there are probably dozens of labs who have worked on characterizing a mammalian version. Isn't it amazing that none of them have noticed any alternative splice variants in spite of the fact that the labs who study splicing have predicted that each one should produce several difference proteins?

  3. Had anyone at the crime labs ever cared to do a proper negativ test (unused cotton swab), what should be a standard procedure anyway, they would have known it was a mistake at the first crime scene. They did not do it, and wasted millions in the most expensive search for a criminal in Germany ever.