40.000 years of human presence in Southern Europe: the Italian case study

The idea that Sicilians and Southern Italians already had ancestry similar to ancient Greeks well before the Greek period starting in the 8th Century BC. comes first of all from common sense (I never bought the idea that the southern Italians were similar to the northern Italians of the Iron Age and then moved further south only thanks to the migrations of the imperial age), from archaeology that assumes deep contacts between southern Italy and the Aegean already in the bronze, and from ancient DNA, see recent studies that have analyzed samples of Sicily of the Bronze Age. The Italian cline was formed definitively after the fall of Rome, when in the Middle Ages the pre-unitary states began to form and Italians gruadually stopped for centuries to move from one part of the country to another. Of course, there continued to be small movements. Like that, for example, of the northern Italians who after the Norman conquest settled in Sicily and Basilicata, or like the Albanians, Balkanites and even Slavs who after the Turkish expansion migrated to Italy, especially in southern Italy.

Agreed, the other idea was more the position of racists, and ironically uber-liberals; who are ever the odd-bedfellows in questions of ethnicity, and history. I guess it takes an impartial and informed mind to see that it is a canard.
 
On the Etruscans they are only defending their old boss, who is an old-fashioned geneticist who has never worked on ancient DNA and who since the late '80s began a work on pre-Roman Italy, already criticized abroad by Robert R. Sokal, which was based on a complete lack of archaeological, anthropological and linguistic knowledge of ancient Italy, and adhered to many theories already discredited by archaeology and even anthropology.





The authors more accurately summarized the situation in Southern Italy and Sicily because they had no old research of their own to defend.

Pax Augusta: Well I was as I alluded to not aware that Professor Piazza is one of these academics who has theories that are so entrenched, he can't allow anything that challenges his world view to be published, etc or he restricts his Doctoral Students from drawing conclusions that contradict his theoretical world view (The Etruscan question). So as you note, since he had no "skin in the game" to use an American Idiom" with respect to Sicilian and Southern Italian DNA, he or his pupils in this case (the research team on this paper) more accurately summarized the situation. So in the context of the Southern Italian and Sicilian situation within the context of ancient DNA and modern DNA questions, I was glad to see this paper make the conclusions it made.
 
Pax Augusta:

Everything starts from these "Synthetic maps of human gene frequencies" that were elaborated between '80s and '90s, and that they tried to demonstrate with subsequent research based on increasing circular arguments. Until the whole-Genome of ancient DNA arrived and further discredited what they were trying to prove aprioristically.

"Synthetic maps of human gene frequencies, which are maps of principal component scores based on correlations of interpolated surfaces, have been popularized widely by L. Cavalli-Sforza, P. Menozzi, and A. Piazza. Such maps are used to make ethnohistorical inferences or to support various demographic or historical hypotheses. We show from first principles and by analyses of real and simulated data that synthetic maps are subject to large errors and that apparent geographic trends may be detected in spatially random data. We conclude that results featured as synthetic maps should be approached with considerable caution."

https://digitalcommons.wayne.edu/humbiol/vol84/iss5/12/
 
Last edited:
Pax Augusta: Well I don't think Principal Components analysis is wrong, it has its strengths and when used correctly is a valid statistical methodology. I was taught it in Graduate School some 25 or so years ago but I don't use it in anything that I do. It seems that Professor Piazza was using data from the 80's and 90's, well before the explosion of ancient DNA and even before the human genome was finally finished and drew inferences based on the data from that time period, which of course was very limited.

So from what you wrote, it seems the issue was more the DNA data they used was more of the problem in that there data that they ran principal components on was what caused incorrect inferences, not Principal components itself. Much of the current DNA research that we read here uses PCA but of course they are using much better DNA samples from a wide range of periods, etc. I went to the link you provided and couldn't get the paper so my statements above are only based on what was presented in the abstract and thus I may be incorrect in my assessment.

Anyway, thanks for the dialogue, Buona giornata.
 
Pax Augusta: Well I don't think Principal Components analysis is wrong, it has its strengths and when used correctly is a valid statistical methodology. I was taught it in Graduate School some 25 or so years ago but I don't use it in anything that I do. It seems that Professor Piazza was using data from the 80's and 90's, well before the explosion of ancient DNA and even before the human genome was finally finished and drew inferences based on the data from that time period, which of course was very limited.

So from what you wrote, it seems the issue was more the DNA data they used was more of the problem in that there data that they ran principal components on was what caused incorrect inferences, not Principal components itself. Much of the current DNA research that we read here uses PCA but of course they are using much better DNA samples from a wide range of periods, etc. I went to the link you provided and couldn't get the paper so my statements above are only based on what was presented in the abstract and thus I may be incorrect in my assessment.

Anyway, thanks for the dialogue, Buona giornata.

The Principal Components analysis can be done with everything, it is not the PCA, a multivariate technique, that has been criticized. Synthetic maps of human gene frequencies have nothing to do with Genome-based Principal Components analysis, Synthetic maps weren't even based on the autosomal DNA of modern samples, they were a completely different stuff created at the end of 70s.
 
Pax Augusta: Well as I said, I only read the abstract. I took Principal Component Scores to mean they ran Principal Components to come up with the smallest number of Factors (components) to explain Populations. So I am not sure then what they mean by Principal components Scores. I took Principal Components analysis in a Multi-variate Statistics course (more economic based area). So when I see Principal Components, I think of running it and getting the eigen values of all the factors and ones that are > 1 are generally included as the variables (explanatory factors) to explain some Dependent variable, etc. Those less than 1 seem to add very little incremental explanatory power, although some might argue 1 is to strict a cut off.

Ok, so I assumed they were using autosomal DNA samples. Again, I only have read the abstract. So if I may ask, can you briefly describe how these Principal Component scores and Synthetic maps works. They used gene frequencies? Ok, well that does seem questionable. Am I to understand they are using Principal Component Scores to be something different than what I described above?
 
Pax Augusta: Well as I said, I only read the abstract. I took Principal Component Scores to mean they ran Principal Components to come up with the smallest number of Factors (components) to explain Populations. So I am not sure then what they mean by Principal components Scores. I took Principal Components analysis in a Multi-variate Statistics course (more economic based area). So when I see Principal Components, I think of running it and getting the eigen values of all the factors and ones that are > 1 are generally included as the variables (explanatory factors) to explain some Dependent variable, etc. Those less than 1 seem to add very little incremental explanatory power, although some might argue 1 is to strict a cut off.

Ok, so I assumed they were using autosomal DNA samples. Again, I only have read the abstract. So if I may ask, can you briefly describe how these Principal Component scores and Synthetic maps works. They used gene frequencies? Ok, well that does seem questionable. Am I to understand they are using Principal Component Scores to be something different than what I described above?


I have read the study. At that time, autosomal DNA analysis did not yet exist. Yes, they used some kind of alleles frequencies. However if you want to understand what I'm saying, and it's me who can't explain myself, read the first introductory chapter of David Reich's book ("Who We Are and How We Got Here") where he talks about Cavalli-Sforza (Piazza was one of the main collaborators of Cavalli-Sforza) and where Reich explains why many things hypothesized by Cavalli-Sforza turned out not to be true, despite acknowledging his role as a pioneer.
 
Pax: Ok, well I would think some of these Researchers would have had access to some DNA even back then, but yes it wasn't the SNP type testing we have today. FBI labs in the USA by the 1980's had DNA and were analyzing it for Crime cases, but I have to go back and see exactly in what context DNA was being used that way. That was a long time ago and I think it related to what was referred to then a Blood test, etc where they had a suspect and sampled his/her relatives and then got matches and arrested their guy.

I have not read Reich's book and I agree I do need to get a copy of it.

Thanks again.
 

This thread has been viewed 8685 times.

Back
Top