Dataset Viewer
Auto-converted to Parquet Duplicate
sample
stringclasses
1 value
type
stringclasses
2 values
PC1
float64
0.01
0.11
PC2
float64
0.01
0.08
PC3
float64
-0.03
-0.01
PC4
float64
-0.03
-0.01
PC5
float64
-0.02
-0.01
PC6
float64
-0.01
-0
PC7
float64
-0
-0
PC8
float64
-0
-0
PC9
float64
-0.01
-0.01
PC10
float64
0
0
PC11
float64
0
0
PC12
float64
0
0
PC13
float64
-0
-0
PC14
float64
0
0
PC15
float64
-0.01
-0.01
PC16
float64
0
0
PC17
float64
0
0
PC18
float64
-0
-0
PC19
float64
0
0
PC20
float64
-0
-0
PC21
float64
-0
-0
PC22
float64
-0
-0
PC23
float64
-0
-0
PC24
float64
0
0
PC25
float64
-0
-0
umay_samli
scaled
0.108132
0.07718
-0.025267
-0.034238
-0.022158
-0.008088
-0.00235
-0.003461
-0.014317
0.002187
0.001137
0.001948
-0.003419
0.001239
-0.007329
0.001326
0.002347
-0.0019
0.00088
-0.002751
-0.002246
-0.000124
-0.000493
0.001807
-0.000838
umay_samli
raw
0.0095
0.0076
-0.0067
-0.0106
-0.0072
-0.0029
-0.001
-0.0015
-0.007
0.0012
0.0007
0.0013
-0.0023
0.0009
-0.0054
0.001
0.0018
-0.0015
0.0007
-0.0022
-0.0018
-0.0001
-0.0004
0.0015
-0.0007

My Personal DNA Results

My personal DNA ancestry results: Global25 (G25) PCA coordinates, GEDmatch calculator runs (MDLP, Eurogenes, Decoded, HarappaWorld, puntDNAL, Gedrosia) and IllustrativeDNA period breakdowns.

This is a personal data archive shared for educational purposes. Admixture calculators are statistical estimates, not clinical or genealogical proof.

Deep Dive Video

Deep Dive of My DNA Results (YouTube)
▶ Click the image to watch the deep dive on YouTube

Repository Outputs

File/directory Content Format
data/g25_coordinates.parquet G25 PCA coordinates, scaled and raw (25 PCs) Parquet (viewer default)
data/admixture_results.parquet Admixture percentages from all 19 calculators (186 rows) Parquet (viewer config admixture)
data/population_shares.parquet Single and mixed-mode population sharing from all calculators (760 rows) Parquet (viewer config population_shares)
data/g25_coordinates.csv Same G25 coordinates, human-readable CSV
results/g25.md G25 coordinates in original text form Markdown
results/mdlp.md MDLP: K11 Modern, K16 Modern, K23b, World-22, World Markdown
results/eurogenes.md Eurogenes: K13, EUtest V2 K15, ANE K7, K9b–K12b, K36, HG vs Farmer, Jtest, EU Test Markdown
results/decoded.md Decoded: V3, World 9, K7b, K12b Markdown
results/harappaworld.md HarappaWorld calculator Markdown
results/puntdnal.md puntDNAL: K10 Ancient, K12 Ancient/Modern, K13 Global, K15 Markdown
results/gedrosia.md Gedrosia: K3, K12, Ancient Eurasia K6 Markdown
results/illustrative_dna.md IllustrativeDNA coordinates and period fits Markdown
images/ PCA plot screenshots and video thumbnail PNG/JPG
raw_notes.md Original unedited notes, kept for provenance Markdown

G25 Coordinates

data/g25_coordinates.parquet is the machine-readable core of the dataset and the config shown in the Hugging Face viewer:

  • sample — sample name (umay_samli)
  • typescaled (recommended for modeling) or raw
  • PC1PC25 — Global25 principal components (float64)

The coordinates can be used directly with standard G25 tooling (Vahaduo, nMonte, PCA plotting).

Admixture Results

data/admixture_results.parquet (viewer config admixture) holds the "Admix Results" table of every calculator in a single machine-readable file (186 rows):

  • tool — source tool (mdlp, eurogenes, decoded, harappaworld, puntdnal, gedrosia)
  • calculator — calculator name (e.g. K11 Modern, World-22)
  • rank — component rank within the calculator result
  • population — population component (e.g. EHG, Caucasian)
  • percent — estimated percentage (float64)

Example: HarappaWorld ranks Caucasian first at 35.18%, followed by Baloch at 17.62%.

Population Shares

data/population_shares.parquet (viewer config population_shares) holds the "Single Population Sharing" and "Mixed Mode Population Sharing" tables of every calculator in a single machine-readable file (760 rows):

  • tool — source tool (mdlp, eurogenes, decoded, harappaworld, puntdnal, gedrosia)
  • calculator — calculator name (e.g. K11 Modern, World-22)
  • modesingle (closest reference populations) or mixed (best two-way fits)
  • rank — rank within the calculator result
  • primary_population — reference population (single mode) or primary mix component (mixed mode)
  • primary_percent — mix percentage for the primary population (mixed mode only)
  • secondary_population, secondary_percent — second mix component (mixed mode only)
  • distance — fit distance (float64)

Example: Possible two-way fits involving a Turkmen reference include Turkmen + Ashkenazi Jewish, Sephardic Jewish, Sicilian, Greek, or Cypriot populations. One of the closest is Decoded K7b's 56.2% Turkmens (Yunusbayev) + 43.8% Ashkenazy_Jews (Behar) at distance 2.92.

Calculator Results

Each file in results/ contains, per calculator:

  • Admix Results — population components with percentages
  • Single Population Sharing — closest reference populations by distance
  • Mixed Mode Population Sharing — best two-way population fits

IllustrativeDNA

results/illustrative_dna.md holds the IllustrativeDNA datasheet and period breakdowns: Bronze Age, Iron Age, Late Antiquity, Middle Ages, plus the closest ancient and modern samples.

PCA Plots

Modern populations (my position marked "YOU"):

PCA plot against modern populations

Historical periods (Ottoman Anatolia, Roman/Byzantine Anatolia, Crimean Tatar, ...):

PCA plot against historical periods

Source and License

  • Source tools: GEDmatch calculators, Global25, IllustrativeDNA
  • License: CC BY 4.0
Downloads last month
62