randman writes:
when we view such a trait in 2 different species and say it cannot occur through convergent evolution because the environment cannot select for a non-adaptive trait, and then assume common ancestry, the assumption could be wrong because the impetus of convergent DNA could well have caused the trait to independently emerge.
This could, of course, happen with just random mutation as well, but is less likely. The introduction of the concept of convergent DNA makes this more likely, imo.
Hi Randman,
If we assume that genetic diversity is generated by random mutations, then we can use simple bayesian theory to calculate the probability that two sequences converged without common ancestry. We could compare this probability to the probability that the two sequences are similar because of common ancestry, and see which hypothesis fits the data best. This is done all the time in molecular evolutionary biology, so your point is basically correct. This kind of convergence is often considered "noise" in the data and the level of noise in each DNA sequence will determine which DNA sequence we use in our analysis.
I wasn't really sure what you meant by "the impetus of convergent DNA". Do you mean that sequences might mutate in a particular direction, so that they are more likely to end up similar by convergence that we would expect from random mutation alone? If this is the case, we would just have to work out what the rules of the impetus are, and factor them in to our probability calculations. For example we know that transition mutations (A->G, G->A, T->C, C->T) are more common that transversion mutations (G->C, A->T, etc.) because of the molecular structure of the nucleotides. We can use this information to make our probability models of sequence evolution better.
Hope this helps,
Mick
This message has been edited by mick, 06-06-2005 11:23 AM