How do we apply evidence of Common Descent?

[Dan_Eastwood]

@pnelson

I think most biologists who haven’t looked into the matter would be stunned by the conflicts in mitochondrial data revealed in the Richards et al. study.

Citations are helpful, rather than having people chase thru videos for the actual source.

[Dan_Eastwood]

I actually know enough about paternity testing to be fairly confident BLAST has nothing to do with it, UNLESS it corresponds to one of the common markers used.

[pnelson (Paul A Nelson)]

Variation Across Mitochondrial Gene Trees Provides Evidence for Systematic...

The pdf will be behind a paywall for anyone without a subscription, so please contact me privately if you’d like to see the full paper (nelsonpa@alumni.uchicago.edu).

[Mung]

@Dan_Eastwood

I actually know enough about paternity testing to be fairly confident BLAST has nothing to do with it, UNLESS it corresponds to one of the common markers used.

I was hoping that was something we could all agree on.

Is the KALIGN program used in paternity testing?

ERVs?

I mean if all humans share the same ERVs, can’t we look at their ERVs and create a phylogeny of the human family tree? Perhaps you and I are more closely related than either of us would be comfortable with.

[T_aquaticus]

@Mung

Is the KALIGN program used in paternity testing?

I already explained it in a previous post:

@T_aquaticus

For paternity tests, I believe they use PCR to amplify short tandem repeats and use gel electrophoresis to measure the size of the amplified bands. They then enter the genotype as the size of the DNA and not the sequence. If there are 4 possible alleles at each STR and you have even mixing of STR alleles then you can find just one person out of billions with relatively few STR’s.

I would think that the sequences are assumed to be identical between PCR fragments of equal length when using the same PCR primers.

@Mung

I mean if all humans share the same ERVs, can’t we look at their ERVs and create a phylogeny of the human family tree? Perhaps you and I are more closely related than either of us would be comfortable with.

If there was enough variation in ERV’s this would be possible.

[Mung]

@T_aquaticus

We can readily observe that cousins and siblings share orthologous ERV’s through common ancestry.

I’m not sure what cousins and siblings have to do with it. Won’t any two humans share orthologous ERV’s through common ancestry?

Are we going to be able to discuss the actual methodologies employed when inferring phlyogenetic relationships and the methodologies employed in paternity testing and compare them?

@T_aquaticus

I am saying that the same concepts that apply for STR’s in paternity testing apply to shared ERV’s among primates.

DNA can be used for comparison purposes. That’s not in dispute.

[Dan_Eastwood]

VERY briefly, given a child and a man who might be the father (a prior assumption that the man and the mother … well … you know), look to see how many common Loci/Markers the man and child have in common. Based on those markers, calculate the probability two people would share those markers at random. Calculate a posterior probability that THIS man is the father.

There are a lot of details I left out, but the basic calculations can be done on a spreadsheet.

More info here: http://dna-view.com/promeg89.htm

[T_aquaticus]

@Mung

I’m not sure what cousins and siblings have to do with it. Won’t any two humans share orthologous ERV’s through common ancestry?

Won’t any two primates share orthologous ERV’s, for the same reason?

@Mung

Are we going to be able to discuss the actual methodologies employed when inferring phlyogenetic relationships and the methodologies employed in paternity testing and compare them?

That’s exactly what I am discussing. The method used for comparing STR alleles is analogous to comparing ERV insertions.

Let’s look at one of the STR sequences:

https://strbase.nist.gov/images/se33.pdf

This section of DNA has multiple repeats of AAAG. Different people have different numbers of these repeats. This means that when you amplify that section the DNA strand will be a different length when run on a gel. They are using the size of the fragment as a proxy for comparing sequence. If there are 4 alleles for this section of DNA in the human population and even mixing then the chances of two people having the same allele is 1 in 4. As you increase the number of STR loci you multiply that probability until it is 1 in several million or billion. If you share a lot of STR alleles then the best explanation is common ancestry instead of random chance.

The same logic applies to ERV’s. Retroviruses insert randomly, so the chances of sharing the same ERV at the same location is small, and the probability nears zero when you look at thousands and hundreds of thousands of ERV’s. As in the case of sharing STR alleles, the low probability of randomly sharing ERV’s supports common descent.

[Mung]

@Dan_Eastwood

More info here: Validation and other Studies

Is that why these cases never go to a jury trial? Could you imagine having to explain that to a jury, lol.

I don’t suppose they use that program or other similar programs for constructing phylogenies.

Can you say more about the “markers” that are used? It’s not just raw DNA comparison, correct?

[Mung]

@T_aquaticus

As in the case of sharing STR alleles, the low probability of randomly sharing ERV’s supports common descent.

Don’t take this personally, it’s not directed at you, but it sounds like a probability argument.

[Dan_Eastwood]

@Mung

Is that why these cases never go to a jury trial? Could you imagine having to explain that to a jury, lol.

I don’t suppose they use that program or other similar programs for constructing phylogenies.

Can you say more about the “markers” that are used? It’s not just raw DNA comparison, correct?

It starts with some really exciting reading of … PROBABILITY TABLES!

No phylogenies or complicated programs that I am aware of. Paternity testing is only interested in the father (rarely, the mother), not cousins or distant relatives.

When these tests come back positive, they are almost foolproof, so this question is unlikely to go to trial. Identical twins can cause problems, but only if both are potential fathers. Sometimes genetic weirdness can happen (like 3 copies of DNA instead of 2), which can change the correct calculations, or sometimes remove all doubt.

[T_aquaticus]

@Mung

Don’t take this personally, it’s not directed at you, but it sounds like a probability argument.

That’s exactly what DNA fingerprinting is. It is a probability argument. That is also what the ERV argument is for common ancestry between humans and chimps.

Let’s say that we are looking at 20 loci with 4 alleles at each loci for DNA fingerprinting. If there is even mixing, the chances of a match at each loci is 1 in 4. The chances of a match at all 20 loci is 4^20, or about 1 in 1 trillion. Since there are only 7 billion people a match between 20 loci is enough to indicate ancestry or the same source if we are talking about forensics at a crime scene.

For retroviruses, the chances of a virus inserting at the same base is about 1 in 50,000 in the most favorable conditions. Therefore, the chances of having more than 99.9% of 200,000 ERV’s at the same location in two genomes is beyond what we would expect by random insertions in two genomes. Therefore, the probabilities favor a single source, or single insertion that occurred in a single common ancestor.

[T_aquaticus]

@Dan_Eastwood

It starts with some really exciting reading of … PROBABILITY TABLES!

Just to make sure I am reading this right . . .

The underlined headers are the loci.

The number under the allele heading is the number of repeats.

The 0.xxx is the percentage of people with a specific allele at a specific loci for a specific population (US African Americans in this case).

Is this correct?

[gbrooks9 (GeorgeB)]

@Mung

Nor does it always lead to murder convictions!

@Mung

Of course. And sometimes it is lacking in phylogenetic conclusiveness.

[gbrooks9 (GeorgeB)]

@Mung

Are you familiar with the claim that if someone accepts the results of a paternity test they ought to accept the evidence for common descent?

But really, the two have nothing to do with each other. Do you agree?

@Mung

A paternity test seeks markers that are so unlikely to be a coincidence that observers can confidently state a man is (or is not) the contributor to an offspring’s genome … or perhaps merely related or not to the father.

This IS an example of testing common descent.

There are different scales involved in testing descent from a population compared to descent from an individual. What makes you think the science radically differs between the two exercises?

[Mung]

@gbrooks9

What makes you think the science radically differs between the two exercises?

I don’t know what you mean by “radically differs.” It’s not a term that i have used. What I have argued is that the methods are different. I’ve also argued that we can test the one using known relationships between fathers and children.

There is no DNA sample that we can take from the “common ancestor” of two extant species.

Can we agree on that?

[colewd (Bill Cole)]

@Mung

There is no DNA sample that we can take from the “common ancestor” of two extant species.

Can we agree on that?

You got my vote.

[Dan_Eastwood]

@T_aquaticus

Just to make sure I am reading this right . . .

The underlined headers are the loci.

The number under the allele heading is the number of repeats.

The 0.xxx is the percentage of people with a specific allele at a specific loci for a specific population (US African Americans in this case).

Is this correct?

That seems right … I made that slide 10 years ago, and my memory is a bit hazy

[Dan_Eastwood]

@Mung

There is no DNA sample that we can take from the “common ancestor” of two extant species.

The Lenski experiments? Finding them in the wild depends on your definition of species.

[glipsnort (Steve Schaffner)]

@Mung

What I have argued is that the methods are different. I’ve also argued that we can test the one using known relationships between fathers and children.

The methods are indeed different, although many of the concepts are the same. For something like tracking closely related flu viruses, on the other hand, the methods can be the same as those used in constructing phylogenies.

@Mung

There is no DNA sample that we can take from the “common ancestor” of two extant species.

Can we agree on that?

Probably true for animals and probably false for plants, which can sometimes speciate at the drop of a proverbial hat (which is, now that I think of it, a really odd phrase).