Which Habsburg Are You?
by Joe Foley, GENE210, 6 August 2010


BACKGROUND
For three centuries, a royal family of Europe engaged in a perverse genetic experiment. The House of Habsburg rose to great power in Austria as Holy Roman Emperors and established a parallel dynasty in Spain. To concentrate their immense holdings within the family, they selectively bred with close relatives, often pairing young nieces with aging, heirless uncles. Thus they also concentrated their deleterious genetic traits, ultimately contributing to the downfall of their Spanish dynasty and a succession crisis that triggered worldwide war.

This program analyzes your SNP genotype from 23andMe to estimate your inbreeding coefficient (F) and place you among the monarchs of history.


USAGE
Preparation: generate a SNP table for the most similar available ancestry using

  script/snp_table.pl <recomb file 1> <recomb file 2> ... <freq file 1> <freq file 2> ... > <output_file>

where recomb files are HapMap genetic maps, downloadable from ftp://ftp.ncbi.nlm.nih.gov/hapmap/recombination/latest/rates/, and freq files are HapMap allele-frequency files, downloadable from ftp://ftp.ncbi.nlm.nih.gov/hapmap/frequencies/latest/forward/non-redundant/ (leave these gzipped). Make sure the recomb files and freq files are in the same order, and use only autosomes.

Running the program:

  ./habsburg.sh <SNP table for your ancestry> <your 23andMe file>

This automates all the substeps, and takes 10-20 minutes.


OUTPUT
You'll be prompted to open an HTML summary of your results (habsburg_results.html), including a short summary of relevant information from Wikipedia. A list of genomic regions inferred to be identical by descent (IBD) will be created in BED format (ibd.bed), for use with the UCSC Genome Browser, along with a list of homozygous 23andMe SNPs within those regions (ibd_snp.txt). FEstim generates additional output that is described in its documentation.


INTERPRETING F
Inbreeding coefficient is the probability that the two alleles at any given locus in an individual's genome are identical by descent, i.e. that both alleles are descended from a single copy in an ancestor who is related to the individual in more than one way. It is equivalent to the kinship coefficient between the individual's parents, and to the inbreeding-related reduction in heterozygosity relative to Hardy-Weinberg equilibrium. Example F values: 0.0625 for the offspring of first cousins, 0.125 for an uncle-niece coupling, 0.25 for brother-sister incest. Inbreeding coefficients are cumulative if there are multiple strange loops in an individual's pedigree.


TECHNICAL DETAILS
This program is a wrapper for FEstim, by Anne-Louise Leutenegger. FEstim estimates inbreeding coefficient from genotype data with a hidden Markov model where each state is the identical-by-descent (IBD) status of a locus, each observation is the genotype at that locus, and each transition probability is a function of the genetic distance between two loci in centiMorgans (in this program, a SNP's position on the genetic map is linearly interpolated from the two nearest SNPs in the reference). Thus, unlike less sophisticated methods, FEstim makes use of the relationship between neighboring loci. However, it fails to account for linkage disequilibrium among dense loci, so this program considers only SNPs that are at least 0.1 cM apart, meaning only a small proportion of SNPs will actually be used. Here, these SNPs are chosen by convenience, not by how representative they are of other loci in their vicinities.

This program is purely for recreational purposes. Results may be imprecise, especially for low values of F.


PRIVACY
The only part of this program that shares private data with another computer is the BED file, if you choose to upload it to UCSC. However, that is a rather anonymous and indirect kind of exposure.


SOURCES
FEstim:
  AL Leutenegger et al. Estimation of the inbreeding coefficient through use of genomic data. (2003) Am. J. Hum. Genet. 73:516-23
  AL Leutenegger et al. Using genomic inbreeding coefficient estimates for homozygosity mapping of rare recessive traits; Application to the Taybi-Linder syndrome. (2006) Am. J. Hum. Genet. 79:62-6
Performance of FEstim:
  O Polasek et al. Comparative assessment of methods for estimating individual genome-wide homozygosity-by-descent from human genomic data. (2010) BMC Genomics 11:139
Habsburg inbreeding coefficients and family tree:
  G Alvarez et al. The Role of Inbreeding in the Extinction of a European Royal Dynasty. (2009) PLoS ONE 4(4)
Biographies and portraits:
  Wikipedia, the free encyclopedia. wikipedia.org

