Figure 1.
A) Nuclear family graph representation of the entire pedigree. Each node represents a nuclear family and families are connected by parent-child relationships. The pedigree includes families with genealogical information but without DNA available (grey), families with SNP genotype data available (white and red) and families with at least one WGS subject (red). B) Visualization by multidimensional scaling (MDS) of population substructure and distribution of affection status in the extended pedigree based on microsatellite genotypes. Individuals with Major Affective disorder (BPI or BPII) are shown in red, unaffected individuals in blue. Subjects with other mental illness are depicted as having unknown phenotype (black). C) Pair wise kinship coefficients for known family relationships in the Amish pedigree and an independent set of unrelated parents from the Autism Genetics Research Exchange (AGRE) [88]. As reported in _ENREF_34 [28], expected kinship coefficients are >0.354 for duplicate samples/monozygotic twins, [0.177–0.354] for 1st degree relatives, [0.0884–0.177] for 2nd degree relative, [0.0442–0.0884] for 3rd degree relatives and <0.0442 for unrelated subjects The expected range of kinship coefficients is shown to the left, the observed kinship coefficients to the right.
Table 1.
Estimates of additive heritability for the BPI, Narrow and BPS phenotype definitions.
Table 2.
Top 30 Amish-specific putative damaging exonic missense variants by prevalence in affected individuals.
Table 3.
Putative damaging exonic variants found in the five linkage regions.
Figure 2.
Comprehensive genetic analysis of bipolar disorder in a genetic isolate.
Pipeline for the integrative analysis of microsatellite, SNP genotype and whole-genome sequence data. Linkage analysis of the ∼2K microsatellite markers is performed to identify genomic regions with evidence for linkage and to identify the nuclear families and specific haplotypes responsible for each linkage signal. Variants segregating on these haplotypes were then identified by phasing and imputation of variants found by whole-genome sequencing. In the final step, family-based association analysis is performed specifically on these variants.
Figure 3.
Linkage peak and haplotype analysis at the 4p16.3 locus.
A) Parametric linkage analysis of chromosome 4 under a recessive model with penetrance parameter 0.85. B) Haplotype frequency distribution in affected and unaffected individuals within the three linked families at 4p16.3.
Figure 4.
Top linkage results from non-parametric and parametric linkage analysis.
Top linkage LOD scores for the three diagnostic schemes BPI, Narrow and BPS from the analysis of the extended pedigree and defined subpedigrees. Suggestive linkage (LOD>2.5) was observed for regions on A) 2p25.3-p25.1, B) 7q21.11-q31.33, C) 16p13.13-13.12 and D) 18p11.22-q12.1. The subset of nuclear families used in the analysis for each peak is shown in red in the nuclear family graph insets. Linkage on chromosomes 7 and 18 were observed using all 49 nuclear families, the peaks on chromosomes 2 and 16 were observed in subpedigrees NB4 and NB6 respectively.