Corsican analysis with Central and North Italians

torzio

Regular Member
Messages
3,945
Reaction score
1,212
Points
113
Location
Eastern Australia
Ethnic group
North East Italian
Y-DNA haplogroup
T1a2 - SK1480
mtDNA haplogroup
H95a
https://www.nature.com/articles/s41598-019-49901-8#MOESM2

[h=1]Genome-wide analysis of Corsican population reveals a close affinity with Northern and Central Italy[/h]
Scientific Reportsvolume 9, Article number: 13581 (2019)


Despite being the fourth largest island in the Mediterranean basin, the genetic variation of Corsica has not been explored as exhaustively as Sardinia, which is situated only 11 km South. However, it is likely that the populations of the two islands shared, at least in part, similar demographic histories. Moreover, the relative small size of the Corsica may have caused genetic isolation, which, in turn, might be relevant under medical and translational perspectives. Here we analysed genome wide data of 16 Corsicans, and integrated with newly (33 individuals) and previously generated samples from West Eurasia and North Africa. Allele frequency, haplotype-based, and ancient genome analyses suggest that although Sardinia and Corsica may have witnessed similar isolation and migration events, the latter is genetically closer to populations from continental Europe, such as Northern and Central Italians.
 
This is the published version of a paper we've already discussed:
https://www.eupedia.com/forum/threads/38967-Corsican-DNA?highlight=Corsican

This was my commentary then:

"Sixteen samples isn't a lot, but they did get some new comparison samples too, and with autosomal analysis you don't need as many samples as for uniparental analysis.

"Corsicansshow an excess of RoHs indicated by the high median values of total number andtotal length of RoHs, next to Sardinians and French Basques (Fig. 1). Thishomozygosity pattern is characteristic of isolated populations with relatively loweffective population size (Ne) and high degree of endogamy29–31."

We suspected the similarity to Tuscans, as the Corsican language bears some similarity to Tuscan.

"To position Corsicans in the context of their geographic neighbours we used anunsupervised clustering approach32 implemented in the program ADMIXTURE33. Atthe level of lowest cross-validation index (K=6), Corsicans and geographically relatedpopulations are composed mainly by three ancestry components - “Sardinian”,“Northern and Eastern Europe”, “Caucasus and Middle East” (dark blue, light blue,and lime green respectively; Fig. 2, Supplementary Fig. S1, S2). Corsicans are mostsimilar to North-Central Italian populations (Piedmont, Lombardy, Tuscany),displaying a slightly larger proportion of a modal component in Sardinians."

"Analysis of average population pairwise Fst distance confirms ADMIXTURE resultsshowing shortest genetic distances between geographically close mainlandpopulations and revealing larger distances with Sardinians and French Basques(Supplementary Table S2)."

F3 shows something different:

"To evaluate the genetic affinity between Corsica and neighbouring populations wecarried out f3-statistics analysis. Outgroup f3 statistics measures the amount ofshared genetic drift between two populations from an outgroup34. The results showedhigh affinity of Corsicans with Sardinians and French Basques, followed, as with theADMIXTURE and FST analyses, by Northern Italians and a series of mainlandEuropean populations (Supplementary Fig. S3A). The tests of admixture based onthe f3 statistics revealed an overall complexity in the admixture history of Corsica.Out of 2,652 tests performed, 18 were statistically significant (Supplementary Fig.S3B), and most of them described Corsica as a combination of Sardinian andCaucasus/Northern European contributions. Moreover, six tests includingpopulations from North African/Arabia and one from Northern Europe werestatistically significant."

"At a finer inspection, out of the 16 Corsican samples analysed, 12 formed apopulation specific cluster (Corsicans2) in the Southern European clade, related toNorth and Central Italian group. Three samples fell into three different clusterscomposed by North and Central Italians (NorthCentralItaly), French (French) andIberians (Iberia2), respectively (Fig. 3, Supplementary Fig. S4), possibly reflecting recent relationship with mainland populations. The remaining Corsican individualformed a separate branch, closest to Portuguese and Spaniards (Iberia1). Sardiniansand French Basques group with neighbouring populations: the former with SouthernEuropeans while the latter with Northern and Western Europeans."

In order to explore the relationships between populations, we performed a PrincipalComponent Analysis (PCA) based on the chunkcount coancestry matrix. The firsttwo PCs explain a large proportion of the total variance (55% and 19% respectively,Fig. 3B), confirming the efficacy in summarizing the genomic information of thepainting approach. The first principal component separates populations along aNorth-South axis, placing North-East Europeans on one side (left) and NearEast/Caucasus populations to the opposite (right). The second component separatespopulations along a West-East axis, with Sardinians and French Basques being clearoutliers. Most Corsican samples group together close to Italian and Spanishpopulations; the four samples that did not group with Corsican main cluster, occupyoutlier positions also in the PCA."

What the heck? Did anyone get to the Supplementary section? They did or didn't do the four grandparent test?

I must say I'm not impressed.

"we performed two different GLOBETROTTER analyses,“full” and “non-local”, as previously reported37,38 (Fig. 4, Supplementary Fig. S7 andTable S4). The “full” analysis considers all samples as possible sources, while “nonlocal” excludes Southern European clusters as donors. In both the analysis, a singleadmixture involving more than one source was identified for the main Corsicancluster. This admixture involved North and West Europe and Levant/North Africansources, and occurred between 37 and 76 generations ago, a time period spanningthe fall of the Roman Empire and the invasions by Barbarians and Saracens.In the “full” analysis, the admixture sources included Sardinians, NorthCentralItalians, Spanish and Sicilians (Supplementary Table S4), possibly suggesting geneflows from continental Europe, while the Levant/North African contribution inferred inthe “non-local” analysis was most probably passed to Corsicans hitchhiking onmainland populations. Similar admixture profiles were observed for two Sardinianclusters, with central time estimates of 59 and 44 generations ago. These resultssuggest that similar admixture episodes affected both the Mediterranean islands. The impact of North Africa and Levant is also evident in the remaining Italian and Iberiansamples, highlighting the wide impact of the event."

Well, doing the analysis with non-Southern European sources from northern Europe was a complete waste of time. There were no north and west Europeans on Corsica at the relevant time.

At the same time, no, any "NA or Levant" ancestry need not have "piggy-backed" with migration from the European mainland. Were they too busy to just even use google? Corsica had its own Carthaginian and Saracen past, in the latter case coming from El Andalus. Plus, how much "Levantine" and "North African" do they think Piedmont and Tuscany could have contributed to the Corsicans when they barely have any themselves, and there wasn't a "mass" migration anyway. (If I'm reading it correctly, their "Event 1 and Event 2" analysis is total bilge. )



"In our analysis, the Corsican samples were characterised by a highancestry of European Early Neolithic (56%), similarly to Italian, Spanish and Balkanpopulations (Fig. 5B). In addition, Corsica harboured a relatively high proportionrelated to Iranian Neolithic (22%), while the contribution of Western and EasternHunter Gatherer (WHG, EHG) was smaller, about 11%. According to previousresearches, a substantial proportion of the EHG and Iranian Neolithic (related toCaucasus Hunter Gatherer, CHG) trace back to Bronze Age movements from theSteppe, however, Iranian Neolithic could have arrived to Western Mediterranean withdifferent migration38–40. This seems to be supported also for Corsica when qpAdmanalysis including Steppe_EMBA (Early Middle Bronze Age) and Iranian Neolithic(Iran_N) were considered in the same analysis evaluating the proportions (AnatoliaNeolithic: 33%, Steppe_EMBA: 19%, Iran Neolithic: 14%, Europe MiddleNeolithic/Chalcolithic: 34%). Compared to Corsicans, French and Spanish sampleswere characterised by a smaller proportion of Iranian Neolithic (13% and 15%,respectively), and a slightly higher contribution from Western Hunter Gatherers (14%and 17%).

Big whoops. In looking at the admixture analysis I'm thinking how they and most geneticists and we amateurs, including me, spend so much time parsing the small differences between groups when in all actuality the differences are so minor. Just look at the block of Southern Europeans. If we zoomed out a bit it would just be a big mass.
0hqV5uN.png
zr9ZKXI.png
[/IMG]


Sorry, guys, I'm just getting a bit bored with all of this, and a bit tired of reading sloppy work. Just as an example, when they say Corsicans have a similarity to the people of Piedmont, what they really mean is that they show a similarity to the people from that sample stupidly labeled Piemonte which is in fact from mountainous Liguria. The names of the villages all have Ligure in them, they speak a Ligurian dialect, and they were only recently incorporated into Piemonte. So, as makes perfect sense given for how many centuries Corsica was ruled by Genova, the Corsicans have an affinity for "Ligurians", not particularly the people of Piemonte. Genova didn't just "rule" Corsica, there was migration from Genova "to" Corsica, as in the famous case of Napoleon's mother's family."

Unless there are changes in the paper, I stand by it.
 
What we have learned in all these years with studies of population genetics. Not only the conclusions of many studies may be wrong, but even the samples they use may not be completely accurate. In the Corsican sample, I think I've seen in previous studies, those who go further north who knows if they are really 100% Corsican or are instead mixed with French or something else.

Are Corsicans most similar to North-Central Italian populations (Piedmont, Lombardy, Tuscany)? It was enough to see the results of the Corsicans on Gedmatch to know it, there was even no need for this study.

Corsicans can be modelled as North-Central Italians mixed with Sardinians. We already knew it that and this study confirms it.



We suspected the similarity to Tuscans, as the Corsican language bears some similarity to Tuscan.


The similarity between Tuscan and Corsican is mainly due to the syntax, for everything else they are different. To the ears of a Tuscan-speaking person someone who speaks Corsican is not speaking Tuscan but he is rather speaking a language more similar to some central-southern dialect of Italy. There are expressions in the Corsican language that recall even some of the dialects of extreme southern Italy. Unfortunately, there are still few studies on the Corsican language and its internal varieties.

Then in Corsica there are at least three linguistic areas, with the southernmost one showing a lot of similarity also with the Sardinian language.

Apart from the first few minutes, the rest of the video is in Corsican language.

 
What we have learned in all these years with studies of population genetics. Not only the conclusions of many studies may be wrong, but even the samples they use may not be completely accurate. In the Corsican sample, I think I've seen in previous studies, those who go further north who knows if they are really 100% Corsican or are instead mixed with French or something else.

Are Corsicans most similar to North-Central Italian populations (Piedmont, Lombardy, Tuscany)? It was enough to see the results of the Corsicans on Gedmatch to know it, there was even no need for this study.

Corsicans can be modelled as North-Central Italians mixed with Sardinians. We already knew it that and this study confirms it.






The similarity between Tuscan and Corsican is mainly due to the syntax, for everything else they are different. To the ears of a Tuscan-speaking person someone who speaks Corsican is not speaking Tuscan but he is rather speaking a language more similar to some central-southern dialect of Italy. There are expressions in the Corsican language that recall even some of the dialects of extreme southern Italy. Unfortunately, there are still few studies on the Corsican language and its internal varieties.

Then in Corsica there are at least three linguistic areas, with the southernmost one showing a lot of similarity also with the Sardinian language.

Apart from the first few minutes, the rest of the video is in Corsican language.



agree

Maybe it was just..... the centuries of Genovese ownership of Corsica that we also have this similar mix.
 

This thread has been viewed 4673 times.

Back
Top