Hemophilia A and B are X-linked bleeding disorders commonly resulting from genetic alterations of the genes, F8 and F9. These genes encode Factor VIII and Factor IX of the coagulation cascade; deficiencies in one of these two factors define hemophilia A and hemophilia B, respectively. Numerous genetic variants in F8 and F9 have been identified.  The majority include single nucleotide variants (SNVs), copy number variants (CNVs), and in the case of F8, large intra-chromosomal inversions. Most CNVs have been found through analyses of targeted regions or techniques that detect differences in gene dosage. However, these methods cannot fully characterize the extent of each CNV as they are unable to detect breakpoints precisely. In this study, our goal was to characterize F8 and F9 CNVs in the My Life, Our Future hemophilia cohort that was recently whole genome sequenced (WGS) by the NHLBI TOPMed program. This cohort was mainly comprised samples from males diagnosed with hemophilia A (N=1,959) or hemophilia B (N=227). Comprehensive analysis of WGS data, revealed 64 CNVs in hemophilia A (57 deletions and 7 duplications) and 24 deletions in hemophilia B subjects. In hemophilia A, CNV size ranged from <100bp to > 100Kb and comprised single exon F8 deletions to multi-genic CNVs. Several F8 CNVs (N=14) included the first or last exon and exhibited heterogeneous breakpoints that differentially affected neighboring genes. In hemophilia B, CNV size ranged from approximately 1Kb to 6.5Mb and most included deletions of multiple exons or multi-genic deletions with substantial breakpoint heterogeneity. Most CNVs detected are causal for severe hemophilia (Factor VIII and Factor IX levels < 1%) and a subset of CNVs (N=16) are likely novel causal variants as they have not been previously reported. CNVs in two samples co-occurred with likely deleterious variants and a subset of CNVs co-occurred with likely benign missense variants. Together, these analyses demonstrate substantial heterogeneity in CNV size and breakpoint location. The identification of CNVs which impact a hemophilia gene and neighboring gene(s) may have clinical implications and raises the possibility of previously undetected hemophilia-associated syndromes. Moreover, combined analysis of SNVs and fully characterized CNVs will enable increased understanding of hemophilia presentation and complications.

Red blood cell (RBC) measurements are polygenic traits, and genome-wide association studies (GWAS), exome chip and sequencing analyses have identified hundreds of associated genetic variants in European-, Asian-, African- and Hispanic-ancestry populations. However, additional loci remain undiscovered and causal variant(s) at each locus have not been well characterized. We performed whole genome sequencing (WGS) analyses for hemoglobin (HGB), hematocrit (HCT), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), mean corpuscular volume (MCV), RBC count and red cell distribution width (RDW) in a multi-ethnic population from phase 1 of the NHLBI TOPMed Project. 7,490 Amish-, European- and African-ancestry individuals from the Old Order Amish Study, the Framingham Heart Study and the Jackson Heart Study, respectively, were pooled for single variant tests using inverse-normal transformed residuals as outcomes, with adjustment for age, sex, study, relatedness, population structure, and residual heteroscedasticity. We identified nine putatively novel loci reaching genome-wide significance (P<5E-8), including two each for HGB (4q22-rs10030052 and ADAMTS20-rs563075310, MAF=0.001), HCT (4q22-rs10030052 and chrX: 53479481, MAF=0.001), MCH (1p11-rs114421285 and 2q14.1-rs150467776, MAF=0.013 and 0.005, respectively) and MCHC (ELOVL6-rs111995643 and PHACTR-rs141007575, MAF=0.014 and 0.001, respectively), and one each for RBC count (ITPR3-rs555173304, MAF=0.004) and RDW (chr22: 51068271, MAF=0.001). The novel variants showed disparate allele frequencies across European and African ancestral populations, and will require validation in additional independent samples. We also confirmed seven previously reported loci for associations with multiple RBC traits. The lead variant for six (HFE, BYSL, MYB, TMPRSS6, CD36 and G6PD) was the same as the reported index SNP or in strong LD (r2>0.8). The lead variants at ITFG3/LUC7L for MCH, MCHC, MCV and RBC were in strong LD with the reported index SNP while lead variants for HGB and RDW were not. At the HBB locus, the strongest association signal was rs34598529, with novel evidence of associations with MCH, MCV, RBC and RDW (P≤9E-9). This variant lies in a putative erythroid regulatory element and has been identified in African American beta thalassemia patients. In summary, our initial results suggest identification of novel rare RBC variants from WGS analyses of a multi-ethnic population.