Intra French PCA

https://i.ibb.co/yn52cBrj/image.png
https://i.ibb.co/s9vrzFkB/image.png
https://i.ibb.co/gMhmmWSc/image.png
https://i.ibb.co/39gGrkkc/image.png
https://i.ibb.co/dJcHjcrG/image.png

POPGEN Project: Methodology & Analysis

I. Study Design & Methodology (Based on the PDF)

The objective of the POPGEN project is to build a French reference panel of genomes to capture local genetic diversity.

Source Cohort: The study utilizes the CONSTANCES cohort (220,000 volunteers).

Selection Process:

A questionnaire regarding the birthplaces of parents and grandparents was sent to all volunteers.

Geolocalization & Distance Computation: Participants were selected based on the geographic concentration of their ancestry (using a specific radius $R$) to ensure strong local representation.

Sampling Funnel:

Selection: 15,000 volunteers were selected based on geographic criteria.

Screening: 10,000 individuals underwent genotyping using SNP-chips (GSA).

Final Sequencing: A subset of 4,000 individuals was selected for Whole Genome Sequencing (WGS), evenly distributed across departments.

Analytical Tools: The team used ChromPainter and UMAP to analyze haplotypic data and define clusters, as well as AMOVA (Analysis of Molecular Variance) to quantify genetic distances.

II. Factual Results (from the PDF)

Fine-Scale Structure: The analysis identified 55 fine-structure clusters.

Geographic Correlation: The genetic structure reflects administrative geography, specifically the departments.

Variance (AMOVA):

99.866% of the total variance is found within departments (intra-departmental).

Only 0.128% of the variance is explained by the division into 95 departments.

This indicates a very low but statistically significant differentiation between geographic groups.

------------------------------------------------------------------------------------------------------------------------

III. Advanced Interpretation of a friend:

According to the POPGEN study, the French population forms a single, highly homogeneous genetic "cloud" (99.866 Pct internal variance), yet it also displays a distinct micro-geographic structure. This is explained by two overlapping historical layers.

1. The Tool: IBD Segments as a Biological Clock
Researchers use IBD (Identity-by-Descent) segments—identical stretches of DNA inherited from a common ancestor.

Short segments (1.5–3 cM) represent ancient ancestry (thousands of years ago).

Long segments (5–7 cM and above) represent recent ancestry (last several centuries).

2. The Ancient Layer: The "Stirred" Pot (> 2,000 years ago)
The Data: Short IBD segments are shared widely across all regions of France.

The History: During the Iron Age and the Roman era, the population—primarily a highly similar Gaulish/Gallo-Roman base—was mobile. Soldiers, traders, and administrators moved across the territory, effectively "stirring" the genetic pot.

The Result: Ancestry from this period is shared nationally; there were no rigid genetic borders between regions.

3. The Medieval Layer: The "Ice Cube Tray" (~1,500 to 600 years ago)
The Data: Long IBD segments (5–7 cM) are not shared between distant regions. They appear almost exclusively on the "diagonal" of the study's matrices, meaning they are shared only between people from the same area.

The "Crystallization": This regional structure becomes unmistakably clear starting 20 to 30 generations ago (approx. 600–900 years ago).

The History: With the rise of feudalism, mobility collapsed. The serfdom system legally tied peasants to their land. For a millennium, people lived, married, and died within a small radius (the parish or seigneury).

The Result: This sedentary lifestyle acted like an ice cube tray, "freezing" the previously mobile Gaulish DNA into local compartments. This created the regional "clusters" (Bretons, Normans, Provençals) we see today.

4. Reconciling Ancient Gradients and Modern Structure
It is important to distinguish between deep ancestry and recent structure:

Deep Ancestry: An ancient North-South gradient (e.g., Steppe vs. Neolithic Farmer ancestry) has existed since the Iron Age. It is a broad, subtle signal that coexists with the general homogeneity of the French "cloud."

Recent Structure: The fine-scale clustering seen today is a "micro-geographic" layer. It is not caused by massive external invasions (allochthonous elements) but by local endogamy.

Conclusion: You do not need "foreign" elements (this does not mean that there are none, but the allochthonous elements, which are not really significant, do not explain the different intra-French variations, which already existed in the Iron Age) to explain regional differences. A mobile, related group (Gauls) that stops moving and settles into isolated territories for 1,000 years (Feudalism) will naturally "crystallize" into distinct regional groups, even while remaining part of the same genetic family.