Comparisons of Neandertal mtDNA with modern humans and modern chimpanzees

I was perusing and came across this abstract, released on the web March this year: I was wondering if someone was familiar with this study. It would seem to me that exogenous DNA contamination could be a very large problem, and how previous studies may have overcome this problem. This would also seem to cast doubt on the repeatibility of the sequence data between the three known samples of neanderthal mitDNA.

Mammuthus and Sylas,Thank you both very much. I had a strong suspicion that possible contamination or chemical alterations to the ancient DNA was well understood, but I wasn't sure on the specifics. The paper I cited in my post sounded pretty fishy, and it seemed this sort of problem would be easy to control for (especially if modern human DNA is the suspected contaminant). Thanks again, and I bow before your collective expertise , and I will definitely have to check out Mam's refs when I have a chance.

quote:

---

This paper, by the way, is a comparison of Neandertals with contemporaneous modern humans (Cro Magnon).

---

Bingo, this ends the debate in my opinion, along with the other controls you listed. If there were a problem with ancient DNA (non-random PCR mutations) it should have reared it's ugly head in the Cro Magnon PCR's as well.PS. See mssg 39 as well.

quote:

---

Soon after the Cro-Magnon study came out, Alan Cooper and Svante Pbo wrote an op-ed objecting to the study on the grounds that if a sequence from Cro-Magnon is retrieved that is similar to modern human DNA sequences, it cannot be distinguished absolutely from a potential contamination and therefore must be rejected as authentic.

---

Good point. You would think that there would be detectable differences between Cro-Magnon sequence and modern human sequence, such as y-chromosomal mutations as well as mitDNA mutations. Or even better, recent retrotransposon activity:Ovchinnikov I, Troxel AB, Swergold GD. Genomic characterization of recent human LINE-1 insertions: evidence supporting random insertion. Genome Res. 2001 Dec;11(12):2050-8.However, reliably dating recent mutations and insertions could be just as unreliable as the aDNA itself, so maybe we are stuck with the very problem that you mention.One of the controls that Sylas mentioned was large DNA fragments (I am guessing > 5kb). This would seem to be a reliable indication of modern contamination, no matter the DNA source. I would assume that aDNA is around 100 to 500 bp, and anything over 5kb would automatically raise eyebrows (much less intact chromosomal DNA at 25 to 30 kb).I think that the objections raised by Cooper and Paabo are important, but it seems that it can be worked around.

quote:

---

What I was wondering was just how much difference there is between a DNA sample from a 40,000 year old Homo Sapiens and one from today?

---

Figure 2 from the paper Sylas references in his post, I get the following data (approximations from the graph, not raw data).Sample--------------Age-------Difference from modern humans
"old" Homo________25,000 yo___________0.010
Neandertal________29,000 yo___________0.085Difference = average pairwise difference per base (I think).And there is no statistical overlap between the two groups (old homo and neandertal. Given that the samples are separated by 4,000 years I would say you could count this as contemporaneous, or at least close to it. If there were random interbreeding between Cro-Magnon (old homos) and Neandertals then these numbers should be a lot closer. Also, the differences amongst modern humans appears to be the same as differences between modern humans and Cro-Magnon. The figure in Sylas's post has a box on the y-axis. That is the current difference among humans, and it is level with the box at 25 (25,000 years old on the x axis). I am not sure if there were anatomically modern humans 40,000 years ago, but I think the study cited here might help clarify things a bit.Added in edit: Sylas's post = Message 37 in this thread. He has a hyperlink to the full paper. Figure 2 from that paper is what I am referencing.[This message has been edited Loudmouth, 05-04-2004] {AdminSylas makes a link. By writing [msg=-37] you give a link to a message in your thread. See ubbcode for full description.}This message has been edited by AdminSylas, 05-05-2004 12:54 AM

quote:

---

Preliminary analysis by nucleic acid hybridization indicated that only a small fraction of DNA isolated from mammoth tissue (2-5%) was homologous to DNA of Asian elephant, a close living relative of the mammoth.

---

I was somewhat surprised when I read this. I am assuming that you are using high stringency temperatures, and the low G+C content probably also contributes. Of course, I could just be talking out of my ass since I work on the protein side of things. My mol bio knowledge mostly comes from working with the gene jockies in my lab. Anyway, enough off topic tech talk for now.

quote:

---

In addition, since it is usually assumed beforehand that the DNA is degraded, most researchers amplify short (100-300 bp) overlapping fragments and "walk out" to generate a larger sequence. The problem is if you land on a Numt and then build your sequence based on the Numt i.e. design primers that are sure to bind to the first sequence you retrieved, you will end up walking along the Numt and not the mtDNA.

---

I can definitely picture this. Nested primers would create overlapping sequence, but the source would be difficult to discern unless the "walk" ran into genomic, non-Numt sequence. I seem to remember that the mutation rate for Numt is higher than mitDNA, so this could pose a huge problem. If you hypothesize that neandertal mitDNA is identical to Cro-Magnon mitDNA, you would have to show how the neandertal mitDNA is actually Numt while the Cro-Magnon mitDNA is truly mitDNA.I can definitely understand your preference for megafauna, that is unless you use elephants as lab techs.

quote:

---

I would find this very surprising. Mitochondria have much higher mutation rates than nuclei do -- their DNA polymerase lacks proofreading capability that the nuclear polymerases have.

---

Very good point. Still, there are scenarios that could cause problems if numt instead of mitDNA is sequenced. I think we can all agree that mitDNA genomic insertions will experience different selective pressures than the mitDNA pool. Also, the two sequences will accrue different mutations. I would assume that mutations in numt's would be considered neutral for the most part. However, if this was a problem it should have shown up in the modern human vs Cro-Magnon pairwise comparison.A question for the experts. If there were a mix of numt and mitDNA in the PCR, what would the sequencing results look like. From my limited experience, it would seem that the predominant sequence at the start of the PCR would overpower the less predominant sequence. That is, the peaks coming off of the sequencer would reflect the predominant base at each PCR termination. The argument then is that the ratio of mitDNA to numt is high enough to preclude contamination from genomic sequences.But then again: From:Thalmann O, Hebler J, Poinar HN, Paabo S, Vigilant L. Unreliable mtDNA data due to nuclear insertions: a cautionary tale from analysis of humans and other great apes. Mol Ecol. 2004 Feb;13(2):321-35.I guess I am straddling the mitDNA reliability fence. It would seem that even Mammuthus would have to control for numt PCR amplification, but luckily avoids the possibility of modern DNA contamination. For now, I will stick with the theory of neandertal mitDNA divergence from H. sapien, but tentatively so. Perhaps further studies will demonstrate non mitDNA flanking sequence, and therefore evidence numt amplification. Given the fragmented nature of ancient DNA a shotgun library would be impractical, but perhaps narrowing down possible insertion sites through phylogenetic analysis could result in canidates for genomic flanking sequence [1].[1] Bensasson D, Feldman MW, Petrov DA. Rates of DNA duplication and mitochondrial DNA insertion in the human genome. J Mol Evol. 2003 Sep;57(3):343-54.

quote:

---

This is true. However, Numt generation is an ongoing process and a newly inserted Numt could be identical or only slightly different from bona fide mtDNA.

---

Yep, you got me there. Just from my recent readings, numt generation happens quite often, more than often enough to back your assertion (in primates at least).

quote:

---

Biochem Biophys Res Commun. 1998 Mar 27;244(3):877-83.
Evidence that two reports of mtDNA cytochrome c oxidase "mutations" in Alzheimer's disease are based on nDNA pseudogenes of recent evolutionary origin.

---

This is a great cautionary tale for all scientists. Researchers (myself included) can be tempted to latch onto an observation that fits our hypothesis without fully qualifying or testing the reliability of the observation. Mutated cytC would be a tempting path for further research, but all for not given that the mutation happens outside of the mitDNA.

quote:

---

My real problem with the nea/Cro-magnon work is that it is fairly widely assumed (in the aDNA community) that a sequence is only authentic if it looks different and thus testing the hypothesis of intermixing is not possible or falsifiable...that is a pretty big problem.

---

I hear ya brother. My response to the cytC mutations applies here as well. I think we can agree that scientists jumped on the nea mitDNA data right away, but are only now (ie last few years) calling the data into question. This should be a lesson for creationists, that trying to falsify theory can only strengthen it, or completely falsify it, whichever the case. I think the current research into hominid aDNA is a great example of what we always tell skeptical anti-science types, that science really is competitive and no theory goes unchallenged.

quote:

---

By the way for both Sylas and yourself, there is a newly published nea study that anyone can access since it is in the free access journals

​

PLoS Biol. 2004 Mar;2(3):E57. Epub 2004 Mar 16. Related Articles, Links

​

No Evidence of Neandertal mtDNA Contribution to Early Modern Humans.

​

Serre D, Langaney A, Chech M, Teschler-Nicola M, Paunovic M, Mennecier P, Hofreiter M, Possnert G G, Paabo S.

​

Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany.

---

Do you feel that this is a valid control, and that this study adds credibility to the nea and amh mitDNA data? I tend to think it does, but not being an expert in the field . . .And just out of curiosity, are the authors above colleagues of yours? You don't have to sacrifice your anonymity, just wondering due to both you and the above research group are located in Germany and you are both into aDNA.Added in edit:

quote:

---

If you look at the sequenceing figure in this paper, you can see what a mess Numts can cause...it was not possible from DNA alone to determine the correct sequence...in fact, direct sequencing produced exclusively Numt sequence

​

Greenwood AD, Paabo S.
Nuclear insertion sequences of mitochondrial DNA predominate in hair but not in blood of elephants.
Mol Ecol. 1999 Jan;8(1):133-7.

---

I could order the paper, but that would be abusing our grant money And, I'm far to lazy to go to a local uni library. But, just from your description of "messed up" I can picture what the peaks coming off of the sequence would look like (if that is what you are talking about). Anyway, fully understand the technical problems numts create.This message has been edited by Loudmouth, 05-06-2004 12:25 PM

quote:

---

Do you know if I post the journal name volume number and date etc like a normal reference, if I post the figure here does that violate the copyright? If it does not I will scan it and post it.

---

As far as I know, posting selected figures for no profit is considered "fair usage" under international copyright. Of course, you can't post the paper whole cloth, but selected quotes and figures sounds kosher to me.

quote:

---

Percy before I go and see if he can get it in appropriate format....this is a wonderful example of why academic publishing that is not free access is really crap ..all the BMC journals and PLoS are open access and you can access all figures and data...without paying..the author pays to publish instead.

---

Thanks for the effort, and I totally agree with you. Like I said earlier, I could use my journal account at work but this would be a little unethical since this is non-work related. Also, some journals charge both the authors and the subscribers, or charge for online access. They get you coming and going.

quote:

---

Hi Loudmouth,
I'll send a copy of the article snail mail to you when I get back from the conference where I am presenting and my subsequent trying to sit still on a beach in southern France

---

email: thebread05@yahoo.comNothing better than going to meetings on the lab's dime, especially to the Frech Riviera. Last trip for me was on Cape Cod near Martha's Vineyard. I didn't even have to present anything, which let me enjoy the conference even more. Enjoy.

Rick Rose,Mammuthus is a little busy lately, so I'll try and fill in for the time being. He may come back and correct me on a few points if I get them wrong.In our discussions (Mammuthus and I) we have been talking about the Neandertal mitochondrial DNA data. The studies seem to indicate that there are enough differences between neandertal and human mitochondrial DNA is enough to conclude that humans and neandertal's were separate species. This does not rule out the ability to interbreed, but if interbreeding were a common occurence then the mitochondrial DNA's should be more alike.However, Mammuthus has brought up some very good points that cast doubt on the reliability of the neandertal mitochondrial DNA (mitDNA here after). Mitochondria are organelles that carry their own DNA separate from human genomic DNA. It has been hypothesized that mitochondria are actually trapped bacteria that the cell uses for energy production. However, once in a while the mitDNA can be inserted into the human genome (nuclear mitochondrial DNA, numt hereafter). Numt's are then susceptible to mutation, and those mutations will differ from the mutations accruing in the mitDNA. Therefore, the numt and mitDNA will have different DNA sequences. Also, it is hard for scientists to differentiate between mitDNA and numt's when they are amplifying certain sequences.So the problem then becomes this. Are the differences between neandertal and human mitDNA due to the fact that we are amplifying neandertal numt's. That is, are we comparing apples and oranges. Could it be that the neandertal true mitDNA is very close to human mitDNA in sequence, but numt interference is blinding us to the truth. The post that you replied to is a species to species comparison of the occurence of numt's. In apes and humans, the numt insertion rate is very high compared to other species. We would expect neandertals to also have a high number of numt sequences in their genomic DNA.What I feel is a solid control for this proposed problem is the sequencing of mitDNA from anatomically modern humans (amh) that lived close to the same time as neandertal's. If the age of the mitDNA and genomic DNA is leading to the the sequencing of numt's instead of true mitDNA, then this same problem should be seen in amh's as well. The data shows that ancient amh and neandertal mitDNA differ by the same amount as current human mitDNA. Also, amh mitDNA fits nicely within the normal distribution of current human mitDNA. That is, anatomically modern humans 20,000 years ago had the same mitDNA sequences as we humans do today.Mammuthus's other criticism is this. If amh mitDNA sequence showed signifigant differences from current human mitDNA, it might be thrown out in favor of sequences that are closer to current humans. The same thing may not happen with neandertal mitDNA, the differences might be more easily accepted as being real when in fact it might be due to numt.Well, I hope this helps you understand the predicament surrounding the sequencing of ancient hominid mitDNA. Just as a side note, mitDNA is used for sequencing since there are 1,000's of copies of mitDNA compared to genomic DNA. This is because there are numerous mitochondria, each with their own genome, per human cell. It is thought that you will have better chances of amplifying mitDNA since there are a lot more sequences floating around in the organic remains of ancient species.