Shared errors

Over at Pharyngula, PZ Myers has a post which first destroys some creationist misconception, but then, far more interestingly, goes on to interpret a recent peer-reviewed paper on copy number variants, or CNVs. The whole piece is worth reading, but what I think is worth of a little extra attention is the brief point our shared errors.

An architecture does not imply intent or purpose, but they often imply a history. The pattern described — that chimps and humans share some common structural elements in their genomes — is better described as evidence of common ancestry than of well-designed function. An intron, for instance, is a piece of random, usually useless DNA inserted into the middle of the sequence of a gene that must be excised from RNA before it can be used to make a functional protein. It’s a little piece of garbage that must be cleaned up before the gene product can do its job. That a human and chimpanzee gene has identical introns is an example of an architecture, true enough, but it is of a shared error. Some all-knowing god—he seems to be consistently making the same mistake.

Okay, let’s take the recent hoo-hah with Coldplay and Joe Satriani. Basically, Satriani is claiming Coldplay ripped off one of his songs. There is some fairly compelling evidence to this claim, but it is far from airtight. What we have are four of the same chords repeating through parts of the songs, but only three consecutive notes are truly in common. We can potentially call this one a coincidence (especially since this song has been around forever and Satriani is only suing now that after Coldplay has won a slew of awards. Essentially, we see two instances of people creating similar things.

Now let’s consider someone learning the Satriani song. I don’t feel like finding the actual chord progressions, so let’s just say it goes A, B, C, D. The person begins to learn things, but is apparently a horrible musician and substitutes an F# for the C. Okay, fine. So we have a version of the song out there which is now A, B, F#, D. Now let’s say this person has a friend who wants to rip the song off. But instead of listening to the original Satriani version, he listens to the mutated version with the F#. Now we have some evidence of a copycat. It isn’t very strong evidence because there is just one error. In both instances, we have just four chords. But let’s say another error is made further along in the song. A chord in the bridge is misinterpreted by the original person learning the song. And, naturally, the copycat makes the same error. As we go deeper and deeper into errors, we begin to get better and better evidence of a common origin – the friend was learning from the interpreted version of the song, not the Satriani version, because it is unlikely he would make, say, 5 of the same errors as his friend. The chance for coincidence shrinks while the odds of identifying the correct source rise.

The way this is like CNVs is that we are seeing common errors being made again and again – and these errors are present in both human and chimp genomes. Of course, it should be noted that it isn’t entirely clear if these errors were directly inherited from a common ancestor or if it was the hotspot for ‘making’ errors is what was inherited, but at any rate, it is evidence for our common ancestory with the other apes. There are far too many common errors being made to simply file this under ‘God did it’. The evidence says that is – still and again – superfluous.