Hacker News new | ask | show | jobs
by dnautics 4747 days ago
RT doesn't work that way, it requires a specific site near its target gene in order to make the DNA strand. There is off-target activity, but it is almost certainly vanishingly rare.
2 comments

anyway, if you're really so hung up on this, let's drop the viral intermediate and focus just on pseudogenes (instead of viral retropseudogenes), which are exactly the thing I described in my first comment.

http://en.wikipedia.org/wiki/Pseudogene#Processed

so, if you want to demonstrate your claim, you would need to write a script that showed that there was no pseudogene that ever inserted at a nonspecific site. You can't show that. ergo, my proposal is more likely than yours. Further, it's support by evidence- for example, the genome is studded with p53 pseudogenes that reintegrated from cDNA nonspecifically.

"almost certainly vanishingly rare event" * "incredibly high rate of reproduction of viruses" = almost certainty.

See the first 4 chapters of "Biology of Cancer" by Weinberg.

molecular biology can deal with probabilities that are lower than 1/the number of particles in the universe.
again, have you read "The Biology of Cancer", by Weinberg? There is strong evidence for what I'm saying.
Further, I'm not sure what the relationship of probabilities to the number of particles in the universe is. The probability of any given sequence emitted by hmmer as the "highest probability" is tiny (often 10e-50 or better, and for very good matches, 10e-138). So what's your point?
uhm, no? But do you understand the thermodynamics of base pair mismatching in DNA/RNA duplexes and how to convert the kcal/mol into probabilities?
Sure, please see my papers on that subject.

the work is mostly on okazaki fragments, but a major conclusion from my phd thesis was that B-RNA is stuck because of a very low probability event (the breaking of a large number of hbonds simultaneously).

http://scholar.google.com/citations?view_op=view_citation&hl...

http://scholar.google.com/citations?view_op=view_citation&hl...

http://scholar.google.com/citations?view_op=view_citation&hl...

http://scholar.google.com/citations?view_op=view_citation&hl...

Fine, then, a DNA-RNA base pair mismatch incurs approximately 1.5 kcal/mol penalty, which is a 1:10 less likelihood of matching. 3' UTRs of genes are what you need to reverse transcribe (by accident) the mRNA of interest to generate the cDNA you describe. On average they are around 700 bp, the odds of finding an exact match are about 700/4^18 - then a one bp mismatch is 700/4^17 times a 1:10 hit in terms of competitive binding. Two bp mismatch is 700/4^16 times 1:100. Then you have to consider that it's evolutionarily unlikely to have segments that exactly or inexactly match tRNAs because the resulting dsRNA is likely to engage in silencing via the DROSHA mechanism - probably because our bodies like to guess what - get rid of ssRNA viruses. So these numbers are probably at least on the order of 1:10 or even 1:100 or more in the wrong direction.