Announcement

Collapse
No announcement yet.

Inherited percentages

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • thetick
    replied
    Originally posted by Javelin View Post
    Here's another example from Population Finder:

    Father:
    Europe Finnish, French, Orcadian, Romanian, Russian 40.69% ±26.29%
    Europe (Southern European) Sardinian, Tuscan 59.31% ±26.29%

    Mother:
    Europe (Western European) Basque, French, Orcadian, Spanish 94.21% ±6.14%
    Europe Tuscan, Romanian, Sardinian 5.79% ±6.14%

    Child:
    Europe (Western European) Basque, French, Orcadian 87.34% ±9.89%
    Middle East Palestinian, Adygei, Iranian, Jewish, Mozabite, Bedouin, Bedouin South, Druze 12.66% ±9.89%
    Wow such inconstancy with PF. Just more evidence how bad PF is. More proof of the "phantom" Middle Eastern with Europeans is there to pull southeast and does not represent real ME ancestry in genealogical time.
    Last edited by thetick; 11 March 2012, 11:15 PM.

    Leave a comment:


  • Javelin
    replied
    Here's another example from Population Finder:

    Father:
    Europe Finnish, French, Orcadian, Romanian, Russian 40.69% ±26.29%
    Europe (Southern European) Sardinian, Tuscan 59.31% ±26.29%

    Mother:
    Europe (Western European) Basque, French, Orcadian, Spanish 94.21% ±6.14%
    Europe Tuscan, Romanian, Sardinian 5.79% ±6.14%

    Child:
    Europe (Western European) Basque, French, Orcadian 87.34% ±9.89%
    Middle East Palestinian, Adygei, Iranian, Jewish, Mozabite, Bedouin, Bedouin South, Druze 12.66% ±9.89%

    Leave a comment:


  • irenych
    replied
    Thank you for sharing. So it all basically depends on what one is looking for. Specific inheritances of specific blocks with significant differences between siblings in that respect need not affect the larger picture which has them possessing very similar gene subgroup identities.

    Leave a comment:


  • MMaddi
    replied
    Originally posted by irenych View Post
    Thanks. Just out of curiosity: did you run your results through Dienekes' utilities, and get significantly different (or interestingly different) values for your sister and yourself?
    Here are Dodecad results, with my results in the left column, my sister's in the middle and the average for the South_Italian_Sicilian_D reference group in the right column. I'm one of the 9 members of the South_Italian_Sicilian_D reference group.

    Mediterranean 33.6//32.35//32.9
    Far Asian 0.0//0.00//0.0
    Siberian 0.0//0.00//0.1
    North European 16.4//16.25//12.2
    South Asian 0.1//0.08//0.6
    West African 0.0//0.07//0.2
    Caucasus 31.6//30.01//35.1
    Gedrosia 2.7//3.33//3.5
    East African 1.1//0.83//0.7
    SW Asian 12.2//13.43//12.2
    SE Asian 0.0//0.00//0.0
    NW African 2.4//3.66//2.5

    Leave a comment:


  • irenych
    replied
    Thanks. Just out of curiosity: did you run your results through Dienekes' utilities, and get significantly different (or interestingly different) values for your sister and yourself?
    Last edited by irenych; 11 March 2012, 05:14 PM. Reason: sp

    Leave a comment:


  • MMaddi
    replied
    I don't think that there's a good straightforward answer to the original poster's question.

    Theoretically, a child inherits 50%, give or take a few percentage points, of their DNA from each parent. But recombination is too random to know which parent will contribute large segments relatively intact from one of their parents or grandparents.

    As an example, consider a shared segment on chromsome 7 that a 2nd cousin and I have. The common ancestor is either our great-grandfather or great-grandmother. The segment is huge, 121 cM, and covers over half the chromosome. So, this very large segment was passed intact to her grandmother and my grandfather and then again to my father and her father and then finally to us. This illustrates Geneaddict's point in his post above.

    Another factor is that Population Finder only uses a subset of the 710,000 SNPs that Family Finder tests. According to the FAQ for Population Finder - see https://www.familytreedna.com/faq/an...spx?id=22#1046 - only 295,000 of the SNP results determine ethnic/geographic ancestry percentages. These are the "ancestry informative markers" (AIMs), which are found at different levels in different populations. That's how Population Finder is able to come up with percentages, by looking for AIMs that indicate you have ancestry from certain populations.

    How are these AIMs distributed throughout your autosomal DNA? It seems to me that that introduces another level of randomness into the process. Since I received over half of chromosome 7 from a set of great-grandparents, that might skew Population Finder results for me to the deep ancestry of those great-grandparents, if there are many AIMs in that segment. Yet my sister only shares a 50 cM segment in the same location on chromosome 7 with our 2nd cousin. So, perhaps Population Finder would find significant differences in the percentages for me and my sister.

    I think this illustrates how the randomness of recombination makes it very difficult to infer the ancestry percentages of one parent based on Population Finder results of the other parent and a child.

    Leave a comment:


  • GregKiroKHR1bL1
    replied
    It would be nice to have parental A-DNA to compare with other sibling A-DNA . . .

    Leave a comment:


  • irenych
    replied
    Originally posted by Javelin View Post
    Generally yes but I'd be careful assuming too much with this approach. I have seen several examples of children who match a population that neither parent does.

    What's safer to say is, if you treat the aggregate of populations weighted by percentages as a geographical "spot," you can guess at the spot of the other parent. However, there are multiple ways to arrive at this spot. For very mixed people, the spot may require some unpacking (e.g. an admixed African American may end up in North Africa but this is an average of West Africa and Europe, not actual North African ancestry).

    This is especially true with West Asian and West European, which all Europeans have in some amount. Since you only inherit half of each parent's autosome, it is possible to inherit most of the West Asian from one parent, say.
    So according to the 50:50 overall rule, if both parents are say 50% "west european" and 8% "west asian" and the offspring inherits about 7% w.a. from one would this work across the board? Could these uneven contributions result in a offspring with more than 50% w.e.?

    Leave a comment:


  • Javelin
    replied
    Generally yes but I'd be careful assuming too much with this approach. I have seen several examples of children who match a population that neither parent does.

    What's safer to say is, if you treat the aggregate of populations weighted by percentages as a geographical "spot," you can guess at the spot of the other parent. However, there are multiple ways to arrive at this spot. For very mixed people, the spot may require some unpacking (e.g. an admixed African American may end up in North Africa but this is an average of West Africa and Europe, not actual North African ancestry).

    This is especially true with West Asian and West European, which all Europeans have in some amount. Since you only inherit half of each parent's autosome, it is possible to inherit most of the West Asian from one parent, say.

    Leave a comment:


  • Geneadict
    replied
    From parent to child, yes 50/50. However, can not necessarily be extrapolated to grandparents or earlier ancestors as mix due to recombination can vary dramatically and I have several examples where split back to one side's grandparents is closer to 67/33 than 50/50 - thus 33.5/16.5 between the grandparents and grandchild.

    "Average" will be 50/50, but many people will have variety of admixs back to grandparents and earlier generations.

    Leave a comment:


  • tomcat
    replied
    Yes, you can apply the 50-50 rule-of-thumb to ancestry attributions.

    Leave a comment:


  • irenych
    started a topic Inherited percentages

    Inherited percentages

    I'm just wondering if the 50:50 autosomal rule applies more or less equally to the gene identities of the parents. Assuming that one parent is unavailable for testing, and the other has, say, a 33% value (Gedmatch numbers) as to "West European" and, say an 8% value as to "West Asian", while the offspring is 24% "West European" and 6% "West Asian", would one be entitled to deduce that the other parent was about 15% "West European" and 4% "West Asian"? Is that permissible, or would these values depend on other factors?
Working...
X