First off, I never cited the accuracy of dates derived from sequence coalescence as evidence for common ancestry. I would fully agree that this calculations can only give us estimates and can be inaccurate for many different reasons.
What I was talking about was the general patterns we see in the data. For example, introns diverge more quickly than exons. We see many examples of genetic equidistance that match the expected phylogeny. There is an overwhelming signal of lineage specific mutations above the noise of homoplasy and convergent evolution. When we compare genomes we see the expected pattern of transitions, transversions, and CpG substitutions. There are multiple lines of independent evidence in genomes that all point to common descent and evolutionary mechanisms. Even devout christian biologists agree, such as the head of the NIH and former leader of the Human Genome Project:
And all I am saying is that one of the âlines of evidenceâ,i.e divergence times is not a line of evidence. Because of huge differences in calculated times of divergence in each arm of the tree in cases like the human chimp ancestry.
Sure. But I am not sure how these lines of evidence correspond to the kind of evidence given by DNA fingerprinting.
The probability of a DNA fingerprinting being wrong is extremely miniscule . (One in millions)
Whereas a phylogenetic signal has p values in the range of 0.05, that too in comparison to a null hypothesis that assumes a random distribution.
How is this a valid comparison. You seem to be comparing apples and oranges here.
What is the p value⌠I donât see any values in the highlighted portion.
And you still havenât addressed the issue of how comparing a phylogeny signal and a DNA fingerprinting is like comparing apples and oranges⌠do you agree to this?
I have yet to go through these evidence in detail⌠have just finished checking out the case of mutation rates and predictions of divergence being evidence for common descent.
When I get there, I will comment on it.
I couldnât get the calculator to work, but the p value for perfectly congruent trees for just 20 taxa is 1.2x10^-22, so it will be even smaller for 30 taxa. You can check out the calculations here:
I donât agree. I think they are comparable. Matching STR alleles between the DNA at a crime scene and the DNA from a suspect is very comparable to the match between trees based on morphology and trees based on DNA sequence.
You can check out my thread on the pattern of substitution patterns over here:
Why would you need to account for gain of function for this comparison? The phylogenetic signal is based on lineage specific mutations. In other words, a tree-like structure for genetic changes. This works equally well for non-functional DNA, although the signal gets swamped by homoplasy more quickly in neutrally evolving sequence (if I understand my phylogenetics correctly).
It all depends on what you want to claim that the phylogenetic signal is telling you. If you want to claim that there are random mutations that occur in species then that is one bar and this data works.
If you want to claim random genetic change plus drift and adaption along with common descent is responsible for the diversity of life that is a much higher bar.
Then you need to account for how complex functional sequences appear such as the flight feather by known evolutionary mechanisms alone.
The answer is that probably the innovative mechanism is yet to be discovered. The closest mechanism we know of capable of creating these sequences is a mind.
Common descent is a part of it as well. Random mutations paired with common descent will produce a tree-like structure in sequence data. That is what they are measuring.
If the sequence data has a phylogenetic signal then that bar has been met.
If the genetic differences between species is consistent with random mutations, common descent, and selection then those differences have been accounted for.
Thats an assertion and many smart scientists disagree with. The signal can be there simply from mutations that donât affect function. We know there is redundancy in DNA.
Another assertion that is inconsistent with the empirical data if you look at it without prejudice. We see many changes during the history life that require lots of new functional information. No one has demonstrated that significant amounts of functional information can be formed with known evolutionary mechanisms.
Here is a link to a discussion I recently had with Perry Marshall a supporter of the EES. At the bottom he had posted some links I gave him to gpuccioâs work at uncommon descent that show some history of large genetic information changes.
Not generally true. In neutrally evolving sequences, a greater proportion of the sites are free to vary, and thus thereâs actually less homoplasy as a proportion of all changes. Consider protein-coding sequences: third position transitions are quickly saturated, leaving not much else as phylogenetic signal. But in the typical intron, changes are distributed evenly across the sequence, reducing the proportion of multiple hits and allowing for more of the rarer mutations, e.g. transversions. Introns actually have a better signal for most purposes than exons, for that reason.
âMany smart scientistsâ? I also donât see what the existence of neutral mutations and redundancy has to do with phylogenetic signal. Could you explain?
The phylogenetic signal is that demonstration. If the functional information you are talking about came about through random mutations, selection, and other evolutionary mechanisms then we should see a phylogenetic signal, and we do.
Because random change degrades a sequence. DNA is a sequential arrangement of chemicals. In order to create information you need to add information to that sequence. Natural selection can slow the degradation and so can purifying selection but degradation is what you get. This is true mathematically as well as empirically. Non functional space is much larger than functional space. If you read gpuccioâs articles you will see that he empirically validates this.
