XIBD (X chromosome Identity by Descent)

XIBD (X chromosome Identity by Descent) performs pairwise identity by descent (IBD) mapping on the autosomes and the X chromosome. It estimates global relatedness and implements a hidden Markov model (HMM) to infer genomic regions that are identical by descent (IBD) using unphased genotype data. XIBD allows the use of one of two models to accommodate for linkage disequilibrium (LD):

1. Model 1 is based on the HMM implemented in PLINK (Purcell et al. 2007). This model assumes the input data is in linkage equilibrium (LE). LINKDATAGEN (Bahlo and Bromhead, 2009) can be used to select a set of markers in LE.

2. Model 2 is based on the HMM implemented in RELATE (Albrechtsen et al. 2009). This model implicitly accounts for LD using conditional emission probabilities so there is no need to thin the dataset. This model requires the use of PLINK (Purcell et al. 2007) to calculate pairwise LD between markers.

Model 1 and Model 2 have different run times due to the difference in SNP numbers included in the analysis and complexity of the models. Model 1 has a faster run time than model 2 and it accurately estimates IBD probabilities from the relatively small number of input SNPs while Model 2 performs better at detecting small IBD tracts. XIBD requires population allele frequencies and, for model 2 only, haplotype frequencies. We provide the HapMap phase 2 and 3 genotype data (build 19) as reference datasets for the 11 HapMap populations (The international HapMap consortium, 2003). We also allow the user to supply their own reference dataset or calculate the required frequencies from the input dataset.


XIBD is available as an R package as of August 1st, 2016, and can be downloaded from here. Earlier versions of XIBD are no longer supported. HapMap reference data can be downloaded from here. The package is written and maintained by Lyndal Henden (henden.l@wehi.edu.au). Last modified 29/7/2016.