GraphIBD is a free, open-source IBD association testing software for genome-wide association study analysis. GraphIBD requires an IBD detection method such as Beagle FastIBD to run first. Then GraphIBD builds upon the IBD information to test if the IBD segments show association to the traits. This can be a powerful strategy to detect rare variant associations as shown in (Browning and Thompson, Genetics, 2012.)


GraphIBD provides the following methods:
Standard Pairwise method
Pairwise method detects excessive IBD rate in case/case pairs. The standard approach is permutation test (slow).
Fast-Pairwise method
Performs pairwise method much faster using a novel importance sampling approach.
Virtual markers
Build virtual marker based on the degree of the vertices. Virtual markers can be directly used for association testing. The virtual marker testing is more flexible than pairwise method in that quantitative traits can be tested and covariates can be included.


  • GraphIBD version 0.1.0 containing the followings,

    • GraphIBD.jar (java archive package file)

    • example.bgl (example input file: Beagle format)

    • example.bgl.fibd (example input file: Beagle FastIBD output format)

    • README


Java source code files

  • Source codes will be available shortly.

Version/bug info

  • v0.1.0 (2012-12-29) Initial prototype version deployed

User's guide

  • Option summary:

usage: java -jar GraphIBD.jar [options]
    -beagle <beagle_file fibd_file>               Beagle format file with phenotype row and fastIBD
                                                  result file
    -single <ibd_file phenotype_file>             Single marker IBD information files
    -out <file>                                   Output file prefix (default="outfile")
    -help                                         Print help
    -verbose                                      Print verbosely (default=false)
    -seed <INT>                                   Seed for random number generator (default=current
    -pairwise                                     Perform pairwise test (default=false)
    -pw_method <name>                             Pairwise test: testing method (CN, CC)
    -pw_sampling_scheme <name>                    Pairwise test: Sampling scheme (isample, permute)
    -pw_num_sampling <INT>                        Pairwise test: Number of permutation or isampling
    -pw_adaptive_num_sampling <P1 N1 P2 N2 ...>   Pairwise test: Adaptive sampling which, given P
                                                  and N pairs, samples N times if p-value is < P
    -pw_auto_avg_adjust                           Pairwise test: Automatically adjust for genomic
                                                  average using all SNPs in file (default=false)
    -pw_manual_avg_adjust <CASE_ADJ CONT_ADJ>     Pairwise test: Adjust for genomic average of cases
                                                  and controls with these specified values (default=
                                                  0 0)
    -pw_no_pvalue                                 Pairwise test: Do not compute p-values, compute
                                                  just statistics (default=false)
    -vmarker                                      Generate virtual markers (default=false)
    -vm_auto_avg_adjust                           Virtual marker: Automatically adjust for genomic
                                                  average of virtual marker (default=false)
    -vm_manual_avg_adjust <FILE>                  Virtual marker: Adjust for genomic average for
                                                  each marker with the values in the file
  • Example Running Command:

java -jar GraphIBD.jar -beagle example.bgl example.bgl.fibd
  • Output format:

Four columns: RSID, P-value, Case/case IBD rate, Contrasting IBD rate


Buhm Han, Eun Yong Kang, Soumya Raychaudhuri, Paul I. W. de Bakker and Eleazar Eskin, “Fast Pairwise IBD Association Testing in Genome-wide Association Studies”, Bioinformatics (2013).


Buhm Han : buhmhan (AT) ,

Eun Yong Kang: ekang (AT)

Funding information

B.H., E.Y.K. and E.E. are supported by National Science Foundation grants 0513612, 0731455, 0729049 and 0916676, and NIH grants K25-HL080079 and U01-DA024417. B.H. is supported by the Samsung Scholarship. This research was supported in part by the University of California, Los Angeles subcontract of contract N01-ES-45530 from the National Toxicology Program and National Institute of Environmental Health Sciences to Perlegen Sciences.