Page 22 of 22 FirstFirst ... 12202122
Results 211 to 217 of 217

Thread: Finally ready to start testing a new approach to oracles

  1. #211
    Registered Users
    Posts
    479
    Sex
    Location
    USA
    Nationality
    USA
    Y-DNA
    R-U152
    mtDNA
    H2a2a1

    Canada Quebec United Kingdom Romani Ireland Finland
    Quote Originally Posted by slowian88 View Post
    my results eurogenes k13,k15,puntdnal k13, could you generate my origin? I'm Polish
    I'm starting to restrict myself to doing two runs for anyone that asks. Just so that I can focus on some more development and in anticipation of having less time once the semester starts next week. So, I did both of the k13's.

    Eurogenes ended up with a few iterations as large revisions were suggested. I started with 100% Polish, but the oracle made two suggestions to switch in. After doing that, the resulting output was pretty clean; nice cluster of about the same percentages and all ethnicities are near each other.

    Code:
    Starting known populations: Polish=9.3%; South_Polish=50.2%; Ukrainian_Belgorod=40.5%; Initial stdeviation=0.66783
    
    Admixture sorted by match quality (smaller is better)
    
    1.2% of Sakilli taken from: South_Polish	std dev: 0.87144	match quality:1.3583	max diff: 0.48688%
    1.2% of Piramalai taken from: South_Polish	std dev: 0.8713	match quality:1.3626	max diff: 0.49125%
    1.3% of Kol taken from: South_Polish	std dev: 0.87349	match quality:1.3699	max diff: 0.49637%
    1.2% of Dusadh taken from: South_Polish	std dev: 0.88165	match quality:1.3721	max diff: 0.49049%
    1.3% of Chamar taken from: South_Polish	std dev: 0.86308	match quality:1.3732	max diff: 0.51008%
    1.3% of Kanjar taken from: South_Polish	std dev: 0.8791	match quality:1.3742	max diff: 0.49506%
    1.2% of Velamas taken from: South_Polish	std dev: 0.88389	match quality:1.381	max diff: 0.49712%
    1.3% of Uttar_Pradesh taken from: South_Polish	std dev: 0.87571	match quality:1.3812	max diff: 0.50551%
    1.5% of Brahmin_UP taken from: South_Polish	std dev: 0.88809	match quality:1.3821	max diff: 0.494%
    1.3% of North_Kannadi taken from: South_Polish	std dev: 0.87128	match quality:1.3835	max diff: 0.5122%
    Puntdnal gave a clean result right out of the gate:

    Code:
    Starting known populations: Polish=100%; Initial stdeviation=2.4043
    
    Admixture sorted by match quality (smaller is better)
    
    14% of Balkar taken from: Polish	std dev: 0.27776	match quality:0.41474	max diff: 0.13699%
    14.8% of Kumyk taken from: Polish	std dev: 0.30793	match quality:0.44492	max diff: 0.13699%
    14% of Turkish_Aydin taken from: Polish	std dev: 0.3204	match quality:0.44708	max diff: 0.12668%
    12.1% of Azerbaijan_Azeri taken from: Polish	std dev: 0.32999	match quality:0.46698	max diff: 0.13699%
    13.3% of Turkish taken from: Polish	std dev: 0.33991	match quality:0.4671	max diff: 0.12719%
    12.6% of Turkish_Kayseri taken from: Polish	std dev: 0.33999	match quality:0.47697	max diff: 0.13699%
    12.8% of Dagestan_Azeri taken from: Polish	std dev: 0.34135	match quality:0.47834	max diff: 0.13699%
    12.8% of Ossetian taken from: Polish	std dev: 0.34495	match quality:0.48194	max diff: 0.13699%
    13.9% of Tadjik taken from: Polish	std dev: 0.36087	match quality:0.49785	max diff: 0.13699%
    11.9% of Abkhasian taken from: Polish	std dev: 0.367	match quality:0.50399	max diff: 0.13699%
    I'm noting overall that both calculators are indicating a pull vaguely in the Caucus direction.

  2. The Following User Says Thank You to therrien.joel For This Useful Post:

     slowian88 (01-19-2018)

  3. #212
    Registered Users
    Posts
    479
    Sex
    Location
    USA
    Nationality
    USA
    Y-DNA
    R-U152
    mtDNA
    H2a2a1

    Canada Quebec United Kingdom Romani Ireland Finland
    Quote Originally Posted by Judith View Post
    I have only just found this thread if you need any more British references feel free to use me GEDmatch M730002
    Wow! Someone who actually triggered the noise threshold that I implemented. I was starting to wonder if it actually was working or not.

    This is using Eurogenes K13:

    Code:
    Starting known populations: Southeast_English=50%; Southwest_English=50%; Initial stdeviation=1.5897
    
    Admixture sorted by match quality (smaller is better)
    
    2.3% of French_Basque taken from: Southwest_English	std dev: 0.99399	match quality:2.7343	max diff: 1.7403%
    0% of Abhkasian taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    0% of Adygei taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    0% of Afghan_Pashtun taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    0% of Afghan_Tadjik taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    0% of Afghan_Turkmen taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    0% of Aghan_Hazara taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    0% of Algerian taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    0% of Algerian_Jewish taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    0% of Altaian taken from: Southeast_English	std dev: 1	match quality:2.7857	max diff: 1.7857%
    All those 0% really mean exactly that. I just force the oracle to give me the top 10 matches. But you only scored one suggestion. Now let me explain this a little since that hasn't happened previously. Over the past few days, I identified what appears to be a rule regarding how much variability there is in the population percentages for all ethnicities. That enabled me to account for that variability in the oracle. What that means is that you and your fellow countrymen should not all score exactly the same percentages on a given calculator. Standard oracles might grab onto that variability and assign it to another ethnicity if there is one that happens to match the specific variation from average that you have. But my oracle will determine the expected variability in a population and if you fall within that range it will not attempt to adjust your admixture. It looks like it took a crack at French Basque, but the fit is not good, so even that one is questionable...

    So, congratulations. You are indeed 100% British

  4. #213
    Registered Users
    Posts
    479
    Sex
    Location
    USA
    Nationality
    USA
    Y-DNA
    R-U152
    mtDNA
    H2a2a1

    Canada Quebec United Kingdom Romani Ireland Finland
    Quote Originally Posted by Haize View Post
    Hi! I've only got Basque ancestry and that is the only ethnicity I'm aware of. Here are my results in Eurogenes EUtest:
    Thank you for including your kit#; the data you provided won't work with the oracle. I can only run it off the exact percentages from the calculator, not the output of another oracle. So I grabbed the right data from gedmatch. The oracle at first suggested swapping a large percentage of the Basque with different Spanish ethnicities. I used the highest ranked one and tried that adjustment. This is the result I got after that correction:

    Code:
    Starting known populations: French_Basque=34.9%; Spanish_Aragon=65.1%; Initial stdeviation=0.93468
    
    Admixture sorted by match quality (smaller is better)
    
    5.8% of Orcadian taken from: Spanish_Aragon	std dev: 0.8702	match quality:1.0949	max diff: 0.22471%
    4.7% of West_Norwegian taken from: Spanish_Aragon	std dev: 0.88588	match quality:1.0972	max diff: 0.21135%
    5.6% of West_Scottish taken from: Spanish_Aragon	std dev: 0.89292	match quality:1.1119	max diff: 0.21894%
    6.2% of Southeast_English taken from: Spanish_Aragon	std dev: 0.89436	match quality:1.1148	max diff: 0.22043%
    5.4% of Irish taken from: Spanish_Aragon	std dev: 0.90716	match quality:1.1259	max diff: 0.21877%
    4.3% of Norwegian taken from: Spanish_Aragon	std dev: 0.91313	match quality:1.13	max diff: 0.21684%
    6.5% of Southwest_English taken from: Spanish_Aragon	std dev: 0.89518	match quality:1.1303	max diff: 0.23514%
    4.7% of North_Dutch taken from: Spanish_Aragon	std dev: 0.91812	match quality:1.1444	max diff: 0.22627%
    4.8% of Danish taken from: Spanish_Aragon	std dev: 0.91761	match quality:1.1448	max diff: 0.22719%
    3.6% of Swedish taken from: Spanish_Aragon	std dev: 0.93452	match quality:1.151	max diff: 0.21653%

  5. The Following User Says Thank You to therrien.joel For This Useful Post:

     Haize (01-19-2018)

  6. #214
    Junior Member
    Posts
    6
    Sex
    Ethnicity
    Basque
    Y-DNA
    R1b - DF27
    mtDNA
    H4a1a1a*

    Forgive me if I understand this the wrong way, but what I get is that my ethnicity looks more similar to Spanish-Aragon than to French-Basque. And the deviation from Aragon gives me percentages of Scandinavia and North-Western Europe (Scottish, Irish, English...)

    I know it's a calculator or oracle but I'm a bit confused because I don't have any ancestors from Aragon or from any of those countries. How should I interpret this? Thank you, therrien.joel.

  7. #215
    Registered Users
    Posts
    34
    Sex
    Location
    Poland
    Ethnicity
    Lach
    Nationality
    Polish (Kresowiak)
    Y-DNA
    R1a R-M417
    mtDNA
    T2b

    Poland
    Thanks!! Caucus or Caucasus? So I am Polish with a very small mix of Caucasia or India?

  8. #216
    Registered Users
    Posts
    479
    Sex
    Location
    USA
    Nationality
    USA
    Y-DNA
    R-U152
    mtDNA
    H2a2a1

    Canada Quebec United Kingdom Romani Ireland Finland
    Quote Originally Posted by Haize View Post
    Forgive me if I understand this the wrong way, but what I get is that my ethnicity looks more similar to Spanish-Aragon than to French-Basque. And the deviation from Aragon gives me percentages of Scandinavia and North-Western Europe (Scottish, Irish, English...)

    I know it's a calculator or oracle but I'm a bit confused because I don't have any ancestors from Aragon or from any of those countries. How should I interpret this? Thank you, therrien.joel.
    Yes, for one, relying on one calculator output is asking for trouble. It's best to analyze a few to see if a pattern shows up. Depending on the samples that the calculator's author chose, there may be variances within a modern population that did not get captured correctly. It is also important to remember that this can be a result from much the more distant past. Aragon is right next to the Basque region, so it wouldn't be hugely surprising that there was intermixing over time.

  9. The Following User Says Thank You to therrien.joel For This Useful Post:

     JerryS. (01-20-2018)

  10. #217
    Registered Users
    Posts
    479
    Sex
    Location
    USA
    Nationality
    USA
    Y-DNA
    R-U152
    mtDNA
    H2a2a1

    Canada Quebec United Kingdom Romani Ireland Finland
    Quote Originally Posted by slowian88 View Post
    Thanks!! Caucus or Caucasus? So I am Polish with a very small mix of Caucasia or India?
    That's what it looks like. And yes, Caucasus.

Page 22 of 22 FirstFirst ... 12202122

Similar Threads

  1. Adoption - where to start?
    By flannelberry in forum Inquiries Corner
    Replies: 2
    Last Post: 04-26-2017, 06:47 PM
  2. Replies: 5
    Last Post: 08-28-2016, 04:21 AM
  3. What made you want to start DNA testing?
    By BalkanKiwi in forum General
    Replies: 19
    Last Post: 07-24-2015, 02:24 PM
  4. Replies: 19
    Last Post: 10-24-2013, 06:58 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •