Tuesday, February 22, 2011

Winding down the current submission opportunity

I will end the current submission opportunity in the near future and I will proceed to the next analysis phase of the Project, which will likely include more global Galore analysis, the 3rd update of ADMIXTURE with Dodecad Populations, and, probably, more fine-scale studies of regional samples.

During that time, I will also doublecheck my samples for possible misrepresentations or related samples, so if you want your sample to stay in the Project, and have (willingly or not) misrepresented either your ancestry or your relationship to other samples, a necessary condition is to let me know.

I just want to give a quick update on the status of the Dodecad populations. I currently have 244 individuals in 23 populations, each of which is represented by at least 5 individuals. This is up from 208 individuals/19 populations of a week ago! Thanks Dodecad participants.

In the last week I was able to reach the 5-individual threshold for the French, Chinese, and Japanese, while the Polish and Danish populations are still 1-2 individuals short.

Whatever your background, if you have 4 grandparents from the same European, Asian, or North African ethnic group or country feel free to send me your data while submission is still open, or to contact me if I am likely to accept your data.

9 comments:

  1. Dieneke, will you upgrade to ADMIXTURE version 1.1 and exclude outliers from the reference datasets as Zack of the Harappa Ancestry Project did? Also will you make IBS and ASD genetic distance analyses in addition to Fst and do study on phased data like David of the Eurogenes Genetic Ancestry Project?

    ReplyDelete
  2. Dieneke, will you upgrade to ADMIXTURE version 1.1

    Maybe


    and exclude outliers from the reference datasets as Zack of the Harappa Ancestry Project did?


    The reference datasets are what they are; samples may be excluded for particular purposes in which case it is stated.

    Also will you make IBS and ASD genetic distance analyses in addition to Fst and do study on phased data like David of the Eurogenes Genetic Ancestry Project?

    I'm not particularly interested in chasing after mini segments of DNA as that is more of genealogical than anthropological interest.

    ReplyDelete
  3. The reason why I especially asked about ASD and IBS is that according to David they are both much more accurate than Fst in giving information about true genetic distances.

    ReplyDelete
  4. The reason why I especially asked about ASD and IBS is that according to David they are both much more accurate than Fst in giving information about true genetic distances.

    There are no "true genetic distances" as genetic distances are man-made constructs. The two are also not directly comaprable, as IBS is defined for pairs of individuals, while Fst for pairs of populations.

    ReplyDelete
  5. There are no "true genetic distances" as genetic distances are man-made constructs.

    By the same logic, races too should be man-made constructs; and you would never treat race as an arbitrary concept, would you?

    The two are also not directly comaprable, as IBS is defined for pairs of individuals, while Fst for pairs of populations.

    These are David's words:

    "But I don't think Fst is very reliable, because it's useless for individuals and even for small samples (with the error margin sometimes bigger than the actual Fst scores). Genome-wide IBS similarity and distance scores are much better, especially for individuals vs. individuals, but they also work very well for populations."

    Regarding ASD, he says these:

    "These tools [onur: he means all genetic analysis tools, including the ones like PCA and ADMIXTURE and not just the ones for directly measuring genetic distance] are all good, if the analyses is designed in a robust manner and interpreted correctly. But Fst is the weakest of them all. Also, it probably won't be used for much longer in studies, with Allele Sharing Distance (ASD) now preferred by most people who know their stuff."

    ReplyDelete
  6. These are David's words:

    "But I don't think Fst is very reliable, because it's useless for individuals and even for small samples (with the error margin sometimes bigger than the actual Fst scores). Genome-wide IBS similarity and distance scores are much better, especially for individuals vs. individuals, but they also work very well for populations."


    Not sure what you are arguing about, as I already told you that Fst is only useful for between-population comparisons.

    ReplyDelete
  7. In the above quote David speaks of IBS as a much better alternative to Fst and at the end he adds: "but they also work very well for populations", so he thinks IBS is also useful - very useful - for between-population comparisons and not just for between-individual comparisons.

    ReplyDelete
  8. I'm not particularly interested in chasing after mini segments of DNA as that is more of genealogical than anthropological interest.

    I certainly don't think they are more of genealogical than anthropological interest; they are much more to do with relationships between populations thus with anthropology.

    ReplyDelete
  9. Dieneke, it would be great if you published a list of project participants sorted by ID number giving information about their ethnic backgrounds as Zack does in his project. That would be much more convenient than searching through the ancestry thread.

    ReplyDelete