The world is as it should be in personal genomics

By Razib Khan | April 23, 2012 11:04 pm

I’ve been having some fun with my daughter’s personal genomics. You see, she has her whole pedigree out to r = 1/4. So, for example, contributions from her grandparents seem to be about on this order:

Paternal grandfather = 0.28
Paternal grandmother = 0.22
Maternal grandfather = 0.23
Maternal grandmother = 0.27

I’ve also calculated the number of recombinations which occurred leading up to the gametes which fused to create her. That will be for a future post. But here let’s confirm that she is not inbred. I used plink for this. Here is the description of the command:

Given a large number of SNPs, in a homogeneous sample, it is possible to calculate inbreeding coefficients (i.e. based on the observed versus expected number of homozygous genotypes).

The estimate of F can sometimes be negative. Often this will just reflect random sampling error, but a result that is strongly negative (i.e. an individual has fewer homozygotes than one would expect by chance at the genome-wide level) can reflect other factors, e.g. sample contamination events perhaps.

My main confusion here was which population I should select? Should I select GIH (HapMap Gujaratis?) or CEU (Utah whites)? I ended up on the TSI sample (Tuscans) as a fake compromise. And of course, because she is mixed-race the results came out very negative, as she had way less homozygosity than would be “expected” from the population wide statistic. I also added an inbred friend (his parents are first cousins) as a “control.” Below are two plots which show the result.

My relatively low F should not surprise. South Asians exhibit relatively high levels of total population genetic diversity as far as Eurasians go. Additionally, my recent ancestry is diverse. I have ancestors of various caste backgrounds, as well as some non-South Asians. But my daughter is definitely an outlier. Though do note that my inbred friend is further up the scale of magnitude of effect. The HapMap Tuscans are already outbred, so there’s more room to go “up” than “down.”

CATEGORIZED UNDER: Personal Genomics
  • Nihaya Khateb

    Razib, I will follow your posts until your daughter reaches the age of adolecent. It’s interesting for me to know how will you then analize her behavior? would you find out that part of her behavior is paternal and the other part is maternal? and when you will have another children: who would have mor parts of your brain and who will have more parts of maternal brain? I have read your comparing of behavior between your brothers, and I liked your conclusions because thy resemble mine.But I think that we inherit our brain structure exactly like the way we inherit any other organ. Our behavior depends basicly on the structure of our brain that we inherited from our parents. Because of the complicated function of the brain, scientists do’nt beleive in this idea of mine. I hope that you someday will be convinced with my model for I rely on your observations.

  • Ed

    Very interesting. I didn’t know that the contributions would vary by paternal/maternal grandfathers and grandmothers.

  • Henry Harpending

    Razib a negative F does not necessariy nor even likely represent sample contamination. It is simply the genetic covariance between parents (always relative to some base population.) If the genetic covariance between the parents is negative then the child is (relatively) outbred and F is negative.

    Kinship and inbreeding coefficients are much more general than the pedigree calculations one learns in intro genetics.

    Good looking child by the way. Congrats.


  • Razib Khan

    henry, the authors of plink assume you’re doing GWAS. so they think you’re looking for a very homogeneous population with about equal relatedness. with that in mind, i think they mean to imply that negative F would be a sample you didn’t want your study.

  • Hassan

    Razib, this might sound ignorant/stupid of me. But is there any formal research that shows what the average contributions by each grandparent (the average of paternal grandfather, maternal grandmother’s, etc.)?

    Should I expect a person to generally have paternal grandfather > paternal grandmother and maternal grandmother >maternal grandfather?

  • Razib Khan

    #5, expected value is 0.25. and, you know that for the autosome:

    paternal grandfather + paternal grandmother = 0.5, and the same for maternal.

  • Anthony

    Razib – is there data as to the standard deviation of the contributions of the ancestors?

  • Mark Adams

    Runs of homozygosity (ROH) are also informative of inbreeding and can be applied usefully to individuals. See Kirin et al


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Gene Expression

This blog is about evolution, genetics, genomics and their interstices. Please beware that comments are aggressively moderated. Uncivil or churlish comments will likely get you banned immediately, so make any contribution count!

About Razib Khan

I have degrees in biology and biochemistry, a passion for genetics, history, and philosophy, and shrimp is my favorite food. In relation to nationality I'm a American Northwesterner, in politics I'm a reactionary, and as for religion I have none (I'm an atheist). If you want to know more, see the links at


See More


RSS Razib’s Pinboard

Edifying books

Collapse bottom bar