Is anthrogenica.com gone?

Status
Not open for further replies.
A method to detect recent drift inspired by the same method as G25 has been used in the Danubian limes preprint, with a clear shout-out to Eurogenes in the supplement.

G25 has proven to be a fantastic tool, which makes sense since it's based on Smart PCA. You would have to believe that evil Davidski deliberately engineered G25 to to be extremely accurate for all other populations but not East Mediterranean ones. He put on his devil's horns and screwed with the 1s and 0s so that pesky Near Eastern ancestry would pop up where it wasn't supposed to. But somehow the calculator works for everybody else perfectly. Yeah right.

It’s not that Davidski is evil it’s PCA based coordinates such as G25 are useful for clustering but not for modeling or for that matter the distances are not reliable either. Here’s why. Suppose you are Greek and you want to use PCA or G25 to find your distance to Syrians vs Iberians. Well if 3 people ran a PCA to get G25 coordinates then each G25 will have different coordinates for you and Syrians and Iberians depending on the number of samples in each population and the number of populations.

Each PCA run will show different distances from you to Syrians vs Iberians. PCAs are designed for general visualization of clustering and outliers but they were never intended for modelling or actual distances.

You can start reading these papers to learn why you can’t use G25 or PCA for those types of things

Why most Principal Component Analyses (PCA) in population genetic studies are wrong
https://www.biorxiv.org/content/10.1101/2021.04.11.439381v1.full.pdf

[h=1]Be careful with your principal components
https://onlinelibrary.wiley.com/doi/10.1111/evo.13835

The reason people are not suspicious of G25 is because it correctly shows them close to their population. That’s why they think the modeling and distances must also be correct when in fact they’re not.

I can prove it if you post distances from British to W. Asians and SC Asians vs distances from E. Asians to Siberians

[/h]
 
This site is a cesspool of misinformation. And G25 is actually a professional tool based on the SmartPCA method designed by Nick Patterson and others; it just happens to be used by amateurs. You don't like Eurogenes tools because you don't like what they say. Simple as that.

Yes it’s based on PCA but PCA is not supposed to be used to measure distances or model
 
It’s not that Davidski is evil it’s PCA based coordinates such as G25 are useful for clustering but not for modeling or for that matter the distances are not reliable either. Here’s why. Suppose you are Greek and you want to use PCA or G25 to find your distance to Syrians vs Iberians. Well if 3 people ran a PCA to get G25 coordinates then each G25 will have different coordinates for you and Syrians and Iberians depending on the number of samples in each population and the number of populations.

Each PCA run will show different distances from you to Syrians vs Iberians. PCAs are designed for general visualization of clustering and outliers but they were never intended for modelling or actual distances.

You can start reading these papers to learn why you can’t use G25 or PCA for those types of things


https://www.biorxiv.org/content/10.1101/2021.04.11.439381v1.full.pdf

Be careful with your principal components
https://onlinelibrary.wiley.com/doi/10.1111/evo.13835

The reason people are not suspicious of G25 is because it correctly shows them close to their population. That’s why they think the modeling and distances must also be correct when in fact they’re not.

I can prove it if you post distances from British to W. Asians and SC Asians vs distances from E. Asians to Siberians


PCA is fine but it can be unreliable based on the data you use analyze. So it it is a data problem which is what I alluded to in earlier posts. Now obviously it is one method to use and it should be part of a broad set of methods to analyze the data. Several points. 1) This paper has not been published so I think it has not passed the peer review process. 2) I did a quick literature search on the author and to be fair, he is not well published from what I can tell. Pretty much every paper published by the Reich Harvard Lab and the Paabo/ Krause Max Plank group uses PCA, along with others, as part of their methods to analyze genomes.

The published paper you cite does not criticize PCA, but instructs readers how to do it properly to avoid misinterpretation. spurious correlations, etc.
 
Last edited:
This site is a cesspool of misinformation. And G25 is actually a professional tool based on the SmartPCA method designed by Nick Patterson and others; it just happens to be used by amateurs. You don't like Eurogenes tools because you don't like what they say. Simple as that.
Actually I've never tested with G25, and your buddy made a G25 version of my K8 model that basically show roughly the same results. What I dislike is people using G25 as some kind of gospel inspite of what is confirmed by academic analysis. You have an infraction for provocative behavior, and that's the only warning you're getting. You're on the verge of being banned.
 
A method to detect recent drift inspired by the same method as G25 has been used in the Danubian limes preprint, with a clear shout-out to Eurogenes in the supplement.

G25 has proven to be a fantastic tool, which makes sense since it's based on Smart PCA. You would have to believe that evil Davidski deliberately engineered G25 to to be extremely accurate for all other populations but not East Mediterranean ones. He put on his devil's horns and screwed with the 1s and 0s so that pesky Near Eastern ancestry would pop up where it wasn't supposed to. But somehow the calculator works for everybody else perfectly. Yeah right.

I don't think you read our forum if you think we are hung up on the stuff you're mentioning.
 
I came here to set the record straight on Anthrogenica, which is the subject of this thread. If you want to have a real conversation with me about matters of anthropological/genetic substance, I'm not hard to find. I post on Eurogenes blog, I post on Razib's blog, I post on AG, I'm on Twitter and Reddit. Eupedia is not aligned with my values and will never give what I have to say a fair chance to be discussed reasonably so forget having that discussion with me here. It's not neutral ground. Peace out.
 
PCA is fine but it can be unreliable based on the data you use analyze. So it it is a data problem which is what I alluded to in earlier posts. Now obviously it is one method to use and it should be part of a broad set of methods to analyze the data. Several points. 1) This paper has not been published so I think it has not passed the peer review process. 2) I did a quick literature search on the author and to be fair, he is not well published from what I can tell. Pretty much every paper published by the Reich Harvard Lab and the Paabo/ Krause Max Plank group uses PCA, along with others, as part of their methods to analyze genomes.

PCA is good tool to visualize general relationships of clusters and that’s what the scientific community uses it for. It’s was never meant to be used to measure actual distances or model and no one I know except Davidski uses it for that.

For distances and modeling the scientific community uses qpAdm, FST, IBS or IBD but not PCA. Distances from PCA fluctuate depending on what other samples are in the run therefore not reliable. I remember that Davidski was warned by the guy that created nMonte but he apparently ignored it probably because he was making money selling G25 coordinates. I think he himself resorts to qpadm and not G25 when he wants to be taken seriously.

PCA has its purpose but that purpose doesn’t include distances and modeling like Davidski uses it for. Unfortunately uninformed people assume that since G25 clusters them with their populations then it’s modeling and distances must be correct also but that’s not the case. If someone posts G25 distances British to W. and SC Asians vs E. Asians to Siberians I can show how off G25 is
 
I came here to set the record straight on Anthrogenica, which is the subject of this thread. If you want to have a real conversation with me about matters of anthropological/genetic substance, I'm not hard to find. I post on Eurogenes blog, I post on Razib's blog, I post on AG, I'm on Twitter and Reddit. Eupedia is not aligned with my values and will never give what I have to say a fair chance to be discussed reasonably so forget having that discussion with me here. It's not neutral ground. Peace out.

Well several points:

1) You say you came to support AG. Fine, that is your view. No problem defending what you believe in just as it is no problem for people who don't care for AG. I as I mentioned have never posted there nor read anything there since they do not allow people not signed up to read their forums (which Eupedia does as well but they do give you I think some free reads. at least back when I first signed up). The Apricity allows for people to read what is going on, can't post if not a member and I have lurked their in the past and fine the major focus of the posters there is the Anthropology forum which does the classify me, classify this person, that person, etc, etc, etc using Carelton Coon's taxonomy (which I freely admit I am not a fan of).

2) You also came here and made a case for G25 and in my view made statements that if you can provide citations to support them, I would surely go take a look at them.

3) G25 as I said does a very good job for me with single distances. If you can provide G25 admixture models that align with published papers, please provide them. Take the Lazaridis et al 2014 paper on the ancient Greeks. Run G25 models and compare them to the paper. Do the same for the Ancient Romans in Antonio et al 2019. That is 2 ancient Southern European populations there. You can do the same thing for example for the Bell Beakers from Central and NW Europe. Also the Yamnaya.

So hey I am always looking for different tools to use. I am well aware of how Dodecad 12b performs regarding the Ancient Romans (Antonio et al 2019) and Ancient Greeks (Lazaridis et al 2014) as the results from the Dodecad 12b are consistent with those published papers. I have already clearly stated that the K8 Dodecad 12b model that Jovialis put together is consistent with 3 recent published papers on modern Italian genetics. Jovialis in a post just above said that an individual who is a G25 expert used the same source populations that Jovialis used in his K8 Dodecad12b model and the results were consistent.

That then begs the question, if G25 and Jovialis Dodecad12b model are providing basically the same results using the "same source populations to model modern Italians" which are consistent with those 3 recent papers on modern Italian genetics, that then leads me to think that the statements regarding the ancestral components to model for example Southern Italians or Greeks, etc that are put forth by G25 toadies (this is directed at the toadies, not the G25 model itself) are spurious correlations that are an artifact of samples included in the source data.
 
Last edited:
PCA is good tool to visualize general relationships of clusters and that’s what the scientific community uses it for. It’s was never meant to be used to measure actual distances or model and no one I know except Davidski uses it for that.

For distances and modeling the scientific community uses qpAdm, FST, IBS or IBD but not PCA. Distances from PCA fluctuate depending on what other samples are in the run therefore not reliable. I remember that Davidski was warned by the guy that created nMonte but he apparently ignored it probably because he was making money selling G25 coordinates. I think he himself resorts to qpadm and not G25 when he wants to be taken seriously.

PCA has its purpose but that purpose doesn’t include distances and modeling like Davidski uses it for. Unfortunately uninformed people assume that since G25 clusters them with their populations then it’s modeling and distances must be correct also but that’s not the case. If someone posts G25 distances British to W. and SC Asians vs E. Asians to Siberians I can show how off G25 is

PCA's are used to capture variance in as parsimonious set of variables as possible and you can use them to plot people from population A to B. So yes we agree what its main use is for. I agree it is not for measuring Distances per se. However, you can use Dodecad 12b (the 12 variables in that model) to estimate admixture proportions and then compare them to the published papers which use those methods you mentioned to estimate admixture models. As I noted before the K8 model that Jovialis used to model modern and ancient Roman study from the Antonio et al 2019 align quite well.
 
I see the G25 and 12b as similar in use giving an admixture result on someone or yourself.

The problem is when one thinks these admixture "programs" means what nationality one has , now and in the past............remember that nations are less than 300 years old...........people think nationality goes back to BC times.........this is the issue with the disputes of people across all forums.

Forget about nationality when doing your Admixture .
 
A method to detect recent drift inspired by the same method as G25 has been used in the Danubian limes preprint, with a clear shout-out to Eurogenes in the supplement.

G25 has proven to be a fantastic tool, which makes sense since it's based on Smart PCA. You would have to believe that evil Davidski deliberately engineered G25 to to be extremely accurate for all other populations but not East Mediterranean ones. He put on his devil's horns and screwed with the 1s and 0s so that pesky Near Eastern ancestry would pop up where it wasn't supposed to. But somehow the calculator works for everybody else perfectly. Yeah right.

Knowing the man, I'm sure he made a (not so)secret anti-wog organisation and implemented a in his G25, an algorithm where everyones south of Dunkirk get pesky Near eastern ancestry
 
I came here to set the record straight on Anthrogenica, which is the subject of this thread. If you want to have a real conversation with me about matters of anthropological/genetic substance, I'm not hard to find. I post on Eurogenes blog, I post on Razib's blog, I post on AG, I'm on Twitter and Reddit. Eupedia is not aligned with my values and will never give what I have to say a fair chance to be discussed reasonably so forget having that discussion with me here. It's not neutral ground. Peace out.

Nowhere is neutral ground so why bother about it ?
 
A method to detect recent drift inspired by the same method as G25 has been used in the Danubian limes preprint, with a clear shout-out to Eurogenes in the supplement.

G25 has proven to be a fantastic tool, which makes sense since it's based on Smart PCA. You would have to believe that evil Davidski deliberately engineered G25 to to be extremely accurate for all other populations but not East Mediterranean ones. He put on his devil's horns and screwed with the 1s and 0s so that pesky Near Eastern ancestry would pop up where it wasn't supposed to. But somehow the calculator works for everybody else perfectly. Yeah right.

G25 can be a "fantastic tool" but both the sample selection and treatment (conversion to G25 format) are curated by a guy who has shown an incredible animus towards south Italians and Greeks, motivated by his irrational desire to see them as "un-european" rather than supported by the genetic literature so far, so you see why I dismiss G25 with the same ease Davidsky throws out the window any transparency and "checkability" for his data collection.

On top of that, south Italians have been the most trolled group in anthrofora, and there have been plenty of confirmed cases of fake kits attributed to south Italians (speaking of which, your collection that I saw on anthrogenica seems by far utterly unreliable, but I do not think you care, since truth doesn't support your fantasies), and it is no surprise that Levanticists are eager to take G25 as gospel, that is often Jews or self-proclaimed Jews that have come up with absolutely historically unsound theories as to how ancient Levantines (the part of their ancestry they identify more with) were very important in both hellenistic Greece and Imperial Italy and how actually the vast make up of jewish ancestry formed in situ in the Levant.

We have then: 1) Groups with clear agenda, nordicists and levanticists, curating and using G25; 2) a well renowned history of fake genetic kits attributed to south Italians. So when the results of G25 for (alleged) south Italians clash with the genetic literature, it is very easy to see what is most likely unreliable.

Furthermore, we have neither uniparental nor actual historical (linguistic, archeological and what you will) data that support the east med delusional fantasies, and just to address one of the biggest and most ludicrous historical mistakes I've seen in the ramblings of Erikl86 and company the hellenistic period wasn't an era of "great mixing", it was an era of Greek culture spreading in Anatolia, Iran and even India, Egypt and the Levant, but outside Anatolia Greek culture remained an urban phenomenon, that is the countryside remained mostly local, culturally speaking (even though with Greek influences, and viceversa in the hotspots of Greek culture in the cities), which means that only a minority identified as Greek in those areas; also, Greeks were very xenophobic and most often than not institutionalised segregation between the locals and the Greek colonists. There were occasional Greeks with mixed ancestry travelling and settling around and some exceptions to the xenophobic attitudes of the Greeks (for example in Egypt very wealthy Egyptians could "buy" their Greek status and there was a Greek in Athens that adopted an Ethiopian), but any one that knows something about history will know that history overteems with exceptions, yet the general picture must be reconstructed with the majority of the cases and general tendencies.
 
This site is a cesspool of misinformation. And G25 is actually a professional tool based on the SmartPCA method designed by Nick Patterson and others; it just happens to be used by amateurs. You don't like Eurogenes tools because you don't like what they say. Simple as that.

Well, Basil, I would say that a tool is only as good as the person who created it. G25 may be "based" on Nick Patterson's SmartPCA, but it wasn't created by Nick Patterson. I would also say that if a calculator can't get a reasonable distance for someone with a genealogy going back to the 1500s in a certain area, there's something wrong with the calculator. The same is true if it doesn't produce results consistent with those from multiple academic papers.

As for this site publishing disinformation I would point out that for years, even on dna forums, and then on anthrogenica, I read over and over again that the Yamnaya were master metallurgists who brought it to the west, that they rode on horseback into the west and conquered it with their bronze weapons. To continual and sometimes nasty rebuttal I insisted that the archaeology showed no such thing. Yamnaya metallurgy was initially a very poor copy of the metallurgy in "Old Europe" and Maykop. Corded Ware trundled along in carts, with barely any copper, much less bronze and few horses. The wrist guards they adopted from the farmers. The "conquest" was against a population depleted by famine and further decimated by a plague brought from the steppe. Much of the "story" was a case of the culture of steppe populations of the Late Bronze Age, Iron Age and beyond being back projected into the period of the first movements out of the steppe. For goodness' sakes, we now have proof that the steppe admixed people didn't arrive in Greece until the Middle Bronze Age.

The evidence was all there, but it didn't fit the preferred narrative.

I would also point out that my view wasn't the only one published here, because we argue and disagree amongst ourselves. Some of what I posted was not the consensus here, and in fact was sometimes vigorously opposed, but I will always be grateful to Maciamo for providing a space where people can debate different views without being banned.

The same goes for the issue of the source of the Indo-European languages. We have had vigorous debates here on the matter, but no one gets banned for their opinions.

I'll just add that I haven't been wrong in any of my major predictions yet, and that includes about the Mycenaeans whom I predicted would be mostly "Minoan" like, the Etruscans, whom I thought were most probably autochthonous based on the archaeology and history of them, the fact that the amount of steppe Eurogenes was "modeling" for Northern Indians and even Afghans was absurd, the fact that Corded Ware would have people in it who were mostly EHG and would also show a resurgence of WHG, with quite a bit less of the CHG/Iran Neo of the Yamnaya, that the modern Spaniards would have admixture from the east and North Africa which would move them away from people like the French Basques, that Beaker would have two "versions", one based in Spain and one steppe admixed, and that the steppe admixed Beakers would have more EEF than Corded Ware, that Iran Neo probably entered the Central Mediterranean at least by the Bronze Age or even the Neolithic, that Albanians were not the descendants of Turks, that I2a was not "native" to the Balkans but was brought by Slavs, that Balkan people would have Slavic input, but just as much local EEF and Caucasus like ancestry, and on and on. I won't bore you with all of it.

How much of it did you predict correctly? How much regarding those issues did the people of anthrogenica get right, or Eurogenes, for that matter? That's a rhetorical question, because in almost all of those cases they were wrong and I was right. You can find the proof all over this site. I don't have the power to erase all the posts and threads where I was wrong, unlike people on other sites. Even if I could, I wouldn't do it, because it would be dishonest and dishonorable.

You are better able to see the best possibilities if you closely follow not only the ancient dna, but the archaeology, and where available, the history of the period, keeping in mind that the ancient "historians" were not historians in the modern understanding. It also helps to have common sense and to, as much as humanly possible, remove your own preferences and prejudices and look at it all as objectively as possible.

No honest person could possibly claim that Davidski and many of his supporters on eurogenes don't have an agenda.

Does any of this mean that I won't be proven wrong in some of my predictions? Of course not, but I analyze the papers as honestly and objectively as I can. In the case of the ethnogenesis of the Italian people, I add to that analysis decades of study of the history of the Classical Era. It is clear from the data we have so far that there was a change genetically in Italy from the Republican Era to the Empire and Late Antiquity. The disagreements are over the details. Nor, and I emphasize this, and admit my own preferences, do I find the idea that we absorbed during that time period additional Anatolian Neolithic, and Iran Neolithic, and Levantine Neolithic if that proves to be the case, at all unwelcome.

If I'm proven wrong about some of the details, so be it. I'll admit it. Have any of your friends at anthrogenica admitted they were wrong in any of their predictions?

Well, I'll give you props for one thing. At least you came here to state your case and didn't just say it from your fortress.
 
Great summary of Anthrogenica's failings and narrow-mindedness from Angela.
 
Lately, I never seem to have any juice to upvote. My apologies. I'll revisit certain posts if I get it bac.
 
I came here to set the record straight on Anthrogenica, which is the subject of this thread. If you want to have a real conversation with me about matters of anthropological/genetic substance, I'm not hard to find. I post on Eurogenes blog, I post on Razib's blog, I post on AG, I'm on Twitter and Reddit. Eupedia is not aligned with my values and will never give what I have to say a fair chance to be discussed reasonably so forget having that discussion with me here. It's not neutral ground. Peace out.

I have no problem having those discussions. Nobody denies that Aegean/Anatolia_BA influences had some impact on these populations. I made a model which shows it could vary in individuals. No doubt higher rates of Iberomaurusian indicate North African/Levantine influences. But I find recent papers point to a Mediterranean genetic construct with Minoans as a baseline for analyzing Ancient southern Europeans in the Balkans and southern Italy. That doesn't dismiss other contributions. But I think it shows Minoan + Steppe and/or Italic-like was the formation of southern Italians. Groups on the periphery did add to them too. But I don't think the degree was as formative from them.

The reason is because Minoans perfectly represents the region for the time. Anatolian_N + CHG. Eastern Mediterraneans like the ones found in Antonio et al were modeled mostly as Anatolian_BA which was 35% CHG, 6% Levant_N, and the rest Anatolian_N. These people could of indeed have an impact during Imperial times. But the impact is typically around 20-30% in some south Italians, but also some have very little or none. But these types of people could have already been present during greek colonial times.
 
Status
Not open for further replies.

This thread has been viewed 107214 times.

Back
Top