Upcoming paper on Anglo-Saxon migration period??

[anglesqueville]

Are you aware of any tendency for ADMIXTURE to produce rather extreme, exaggerated results in the direction of one particular source against all others? In their table S4.3 I see a lot of strikingly uniform results. I seem to remember seeing this discussed here or on Eurogenes at some point. I agree the technical aspects of this paper was a bit of a letdown and not quite as up to date with recent innovations in the field as I'd have imagined, although I understand that it's a matter of convenience to some degree when handling such a large number of genomes.
Regarding the f4-stats, you probably saw already that it's based on present day-individuals, so very likely the same cohort of modern Swedes and Norwegians which was used to generate the NOR-component (n=1910). Where those samples are taken from exactly I do not know.

To tell you the truth, the technical aspects of this work give me a very unpleasant impression, in particular this analysis under ADMIXTURE. But I can't say more until I have the authors' genotypes handy.

[Wâldpykjong]

I have asked Ph2ter to make a umap in G vs C format, with the Saxon samples and that of the members here, go and find yourself if you have G vs C coordinates of course!

https://drive.google.com/file/d/13VQ...ew?usp=sharing

Looks interesting Finn!

[DFSTFD]

This is however what the authors of this study do when they calculate a series of F4 with the "Norwegian" and "Swedish" sources (without telling us precisely, unless I missed this precision, which genomes are used ), table 6.2 if I remember correctly. By the way, they set the significance threshold of their stats to |Z|>2, whereas tradition sets it rather to |Z|>3.

On that particular technical point, for which I can't really weigh on myself, I recall some previous studies considering |Z|>2 results at times. An example that comes to mind (the others I can't recall exactly but I have the general sense that this has been the case elsewhere too!) is Lazaridis et al. 2017 - Genetic Origins of the Minoans and Myceneans, from the supplement:

When we apply this intuition to the best models for the Mycenaeans(Extended Data Fig. 6), we observe that none of them clearly outperforms the others as there are no statistics with |Z|>3 (Table S2.28). However, we do notice that the model 79%Minoan_Lasithi+21%Europe_LNBA tends to share more drift with Mycenaeans (at the |Z|>2 level).

Just for what it's worth, it might not be best practice or at least a more tentative even if somewhat significant conclusion though again I couldn't tell you, but it seems that it's considered sometimes by these studies. If that's actually the case, might be worth a question to e.g. Lazaridis or Patterson if someone has communication with them about some specifics.

[anglesqueville]

On that particular technical point, for which I can't really weigh on myself, I recall some previous studies considering |Z|>2 results at times. An example that comes to mind (the others I can't recall exactly but I have the general sense that this has been the case elsewhere too!) is Lazaridis et al. 2017 - Genetic Origins of the Minoans and Myceneans, from the supplement:



Just for what it's worth, it might not be best practice or at least a more tentative even if somewhat significant conclusion though again I couldn't tell you, but it seems that it's considered sometimes by these studies. If that's actually the case, might be worth a question to e.g. Lazaridis or Patterson if someone has communication with them about some specifics.

They are also rather accommodating for the p-values of some qpAdm models. Well, they are professionals, not me. But after gouging my eyes out at the additional information, I can't help but feel some unease.

[DFSTFD]

They are also rather accommodating for the p-values of some qpAdm models. Well, they are professionals, not me. But after gouging my eyes out at the additional information, I can't help but feel some unease.

That's true. I recall e.g. an accepted model with 0.048 p-value from the Olalde et al. (2021) - Cosmopolitanism... paper, which is quite close but still a bit accommodating as you put it. Either way, you raise some interesting questions for us amateurs who are less knowledgeable on the technical part of the area than you are, and so probably tend to ignore more at times even though important, that might do with some overall clarification.

[Dewsloth]

I'm at work right now (its a slow day here if you can believe it ) and I cant connect my Linux (DNA) laptop to the office wireless, but once I get home tonight I'll post what my program outputs for this sample. I'm using google sheets on the work computer and typing commands into my laptop. If I remember right, it didn't give me any confidence that he was DF19+, but I'll let everyone evaluate it for themselves once I get the opportunity

Thanks! I note almost exactly the same thing happened with the Patterson paper and the same SNP popping up as what turned out to be a false positive as an automated SNP hit for sample I16182:

https://anthrogenica.com/showthread....l=1#post812507

It could be that L1199 just isn't a reliable indicator at all. In Alex's L1200+ block, he doesn't use L1199 by that designation, anyway:

https://www.ytree.net/DisplayTree.ph...085&star=false

Where FTDNA calls the block L1200, Alex calls it R-P312/S116 > DF19/S232 > DF88 > Z17260

In the Patterson paper, despite the initial L1199 hit, iirc the I16182 sample was called by FTDNA:
I16182 R-L2 R-M207>M173>M343>L754>L389>P297>M269>L23>L51>P310>L1 51>P312>ZZ11>U152>L2

[PLogan]

To tell you the truth, the technical aspects of this work give me a very unpleasant impression, in particular this analysis under ADMIXTURE. But I can't say more until I have the authors' genotypes handy.

On that point about the genotypes, any word on release and availability?

[Finn]

The Weser-Elbe triangle was in the early migration period in fact the power house. From these area came a major push towards the Adventus Saxonum....

Johan Nicolay:

[Finn]

Hypothesis: the Adventus Saxonum, the spread of CNE, was in fact hop step jump.

Hop: from the bottleneck (Schleswig/Jutland) to the Weser-Elbe triangle.

Step: from the Weser Elbe along the Dutch North Sea Coast

Jump: to the English North Sea Coast and the Channel.

[anglesqueville]

On that point about the genotypes, any word on release and availability?

I've mailed Schiffels today.