The link I quoted says
"Ultimately, THERE IS NO WAY to measure whether a particular phylogenetic hypothesis is accurate or not.."
This is not a case of tentatively it is an acknowledgement of a real conceptual limit. I am talking about whether of homology or relation claim can be logically or validly proven.
It can be as validly proven as any statistical correlation in science. That is what your quote was saying. Can we know with 100% certaintly? No. However, using statistics we can assign a probability to the phylogeny being accurate. A p value of 99% says that there is a 1% chance of a random set of data producing a false positive. When phylogeny after phylogeny has high p values, then we can be quite confident that evolution is accurate.
I think you are clearly overstating the case for the evidence.
If anything, it is being understated.
So a close genetic pattern in families benefits from lots of shared features. Wider genetic similarities are seriously diluted of relevance so that we begin to have similar sequences to a Banana.
The evidence is that the similarities and differences produce a nested hierarchy. Not only that, but we can infer rates of mutation for genes and junk DNA. We can also use genetic markers to track relatedness, such as ERV's. These are one time mutational events that we can use to reconstruct phylogenies.
You wouldn't convict a killer based on the DNA being that of a male or being that of and African male. The DNA has to be actually their DNA to make a claim with that kind or level of ramification.
I think you would be surprised at how DNA is used in forensics. They don't look at the sequence of the DNA. Rather, they compare haplotypes (i.e. alleles) for short stretches of repetitive DNA called STR's (short tandem repeats). They will look at multiple STR loci and see which allele you have in each case, using PCR and a gel to find the the number of repeats for each STR. What you end up with is something akin to a serial number. For example, if there are 4 randomly distributed alleles for STR A, then matching the suspect at this loci will give them a 75% probable match. With each additional match that probability gets higher and higher and higher. It is actually the same type of statistical test that they use to construct phylogenies.
Edited by Taq, : No reason given.
™ Version 4.2
(1)
