Limits...
Genome wide association study of fetal hemoglobin in sickle cell anemia in Tanzania.

Mtatiro SN, Singh T, Rooks H, Mgaya J, Mariki H, Soka D, Mmbando B, Msaki E, Kolder I, Thein SL, Menzel S, Cox SE, Makani J, Barrett JC - PLoS ONE (2014)

Bottom Line: We successfully replicated known associations for HbF near BCL11A and the HBS1L-MYB intergenic polymorphisms (HMIP), including multiple independent effects near BCL11A, consistent with previous reports.This is the largest GWAS study in SCA in Africa.We have confirmed known associations and identified new genetic associations with HbF that require further replication in SCA populations in Africa.

View Article: PubMed Central - PubMed

Affiliation: Muhimbili Wellcome Programme, Department of Haematology and Blood Transfusion, Muhimbili University of Health and Allied Science, Dar-es-salaam, Tanzania; Department of Biological Sciences, Dar es Salaam University College of Education, Dar-es-salaam, Tanzania.

ABSTRACT

Background: Fetal hemoglobin (HbF) is an important modulator of sickle cell disease (SCD). HbF has previously been shown to be affected by variants at three loci on chromosomes 2, 6 and 11, but it is likely that additional loci remain to be discovered.

Methods and findings: We conducted a genome-wide association study (GWAS) in 1,213 SCA (HbSS/HbSβ0) patients in Tanzania. Genotyping was done with Illumina Omni2.5 array and imputation using 1000 Genomes Phase I release data. Association with HbF was analysed using a linear mixed model to control for complex population structure within our study. We successfully replicated known associations for HbF near BCL11A and the HBS1L-MYB intergenic polymorphisms (HMIP), including multiple independent effects near BCL11A, consistent with previous reports. We observed eight additional associations with P<10(-6). These associations could not be replicated in a SCA population in the UK.

Conclusions: This is the largest GWAS study in SCA in Africa. We have confirmed known associations and identified new genetic associations with HbF that require further replication in SCA populations in Africa.

Show MeSH

Related in: MedlinePlus

Regional association plot of the chromosome 19 region with our suggestive novel SNP (rs28567737).The left y axis represents the negative log10 P values of the associations of SNPs in this region. Each SNP is represented by a dot, with grey dots indicating low LD with rs28567737. The right y axis represents the recombination rate (in centimorgans [cM] per megabase [Mb]). Genes found in the region are shown in relative position under the plot; arrows indicate the direction of transcription.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4221031&req=5

pone-0111464-g006: Regional association plot of the chromosome 19 region with our suggestive novel SNP (rs28567737).The left y axis represents the negative log10 P values of the associations of SNPs in this region. Each SNP is represented by a dot, with grey dots indicating low LD with rs28567737. The right y axis represents the recombination rate (in centimorgans [cM] per megabase [Mb]). Genes found in the region are shown in relative position under the plot; arrows indicate the direction of transcription.

Mentions: The discovery of additional loci has become increasingly challenging mostly because the effect size of a new locus is likely to be lower than that of the 3 primary loci. Our study has identified 8 suggestive novel SNPs worthy of follow up in larger studies. One of these, rs28567737 on chromosome 19 (Figure 6), was genome-wide significant (4.9×10−8), and was in high LD (0.991) with rs10414361 (P = 1.31×10−7), a variant directly typed on the Omni2.5 assay. The lack of validation of the suggestive loci in the UK cohort may be due to the small sample size or different ancestry. The latter is likely to be more significant as most of the UK patients have a Caribbean and West African origin while the population in our study is from East Africa. In addition, studies of SCD suggest that there is a considerable heterogeneity in the genetic and environmental composition of SCD populations within Africa [23], [24] let alone in different continents. Based on differences in ancestry, it is likely that these populations would have different sickle haplotypes with the majority of individuals within the UK cohort carrying the Senegal/Benin haplotype while most of the Tanzanian patients would carry the Bantu/Central African Republic (CAR) in. In addition, HbF regulation pathways may develop differently in different populations [8], hence, allele frequencies may differ across populations. Such a difference has been previously observed for HMIP tag SNP (rs9399137) variant which was found to be less common in the Tanzanian (MAF = 0.01) compared with the UK patients (MAF = 0.07) [12]. In this study we observed lower minor allele frequencies for the suggestive variants within the replication cohort compared to the discovery cohort (Table 2). Another difference observed was high levels of HbF in the UK cohort than those of Tanzanian patients. It is therefore critical that replication studies are done with cohorts within the same geographical region as it is likely that there will be more homogeneity in ancestry. The best replication population for our study would have been from East Africa, however, SCA studies with required HbF data and DNA samples are scarce.


Genome wide association study of fetal hemoglobin in sickle cell anemia in Tanzania.

Mtatiro SN, Singh T, Rooks H, Mgaya J, Mariki H, Soka D, Mmbando B, Msaki E, Kolder I, Thein SL, Menzel S, Cox SE, Makani J, Barrett JC - PLoS ONE (2014)

Regional association plot of the chromosome 19 region with our suggestive novel SNP (rs28567737).The left y axis represents the negative log10 P values of the associations of SNPs in this region. Each SNP is represented by a dot, with grey dots indicating low LD with rs28567737. The right y axis represents the recombination rate (in centimorgans [cM] per megabase [Mb]). Genes found in the region are shown in relative position under the plot; arrows indicate the direction of transcription.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4221031&req=5

pone-0111464-g006: Regional association plot of the chromosome 19 region with our suggestive novel SNP (rs28567737).The left y axis represents the negative log10 P values of the associations of SNPs in this region. Each SNP is represented by a dot, with grey dots indicating low LD with rs28567737. The right y axis represents the recombination rate (in centimorgans [cM] per megabase [Mb]). Genes found in the region are shown in relative position under the plot; arrows indicate the direction of transcription.
Mentions: The discovery of additional loci has become increasingly challenging mostly because the effect size of a new locus is likely to be lower than that of the 3 primary loci. Our study has identified 8 suggestive novel SNPs worthy of follow up in larger studies. One of these, rs28567737 on chromosome 19 (Figure 6), was genome-wide significant (4.9×10−8), and was in high LD (0.991) with rs10414361 (P = 1.31×10−7), a variant directly typed on the Omni2.5 assay. The lack of validation of the suggestive loci in the UK cohort may be due to the small sample size or different ancestry. The latter is likely to be more significant as most of the UK patients have a Caribbean and West African origin while the population in our study is from East Africa. In addition, studies of SCD suggest that there is a considerable heterogeneity in the genetic and environmental composition of SCD populations within Africa [23], [24] let alone in different continents. Based on differences in ancestry, it is likely that these populations would have different sickle haplotypes with the majority of individuals within the UK cohort carrying the Senegal/Benin haplotype while most of the Tanzanian patients would carry the Bantu/Central African Republic (CAR) in. In addition, HbF regulation pathways may develop differently in different populations [8], hence, allele frequencies may differ across populations. Such a difference has been previously observed for HMIP tag SNP (rs9399137) variant which was found to be less common in the Tanzanian (MAF = 0.01) compared with the UK patients (MAF = 0.07) [12]. In this study we observed lower minor allele frequencies for the suggestive variants within the replication cohort compared to the discovery cohort (Table 2). Another difference observed was high levels of HbF in the UK cohort than those of Tanzanian patients. It is therefore critical that replication studies are done with cohorts within the same geographical region as it is likely that there will be more homogeneity in ancestry. The best replication population for our study would have been from East Africa, however, SCA studies with required HbF data and DNA samples are scarce.

Bottom Line: We successfully replicated known associations for HbF near BCL11A and the HBS1L-MYB intergenic polymorphisms (HMIP), including multiple independent effects near BCL11A, consistent with previous reports.This is the largest GWAS study in SCA in Africa.We have confirmed known associations and identified new genetic associations with HbF that require further replication in SCA populations in Africa.

View Article: PubMed Central - PubMed

Affiliation: Muhimbili Wellcome Programme, Department of Haematology and Blood Transfusion, Muhimbili University of Health and Allied Science, Dar-es-salaam, Tanzania; Department of Biological Sciences, Dar es Salaam University College of Education, Dar-es-salaam, Tanzania.

ABSTRACT

Background: Fetal hemoglobin (HbF) is an important modulator of sickle cell disease (SCD). HbF has previously been shown to be affected by variants at three loci on chromosomes 2, 6 and 11, but it is likely that additional loci remain to be discovered.

Methods and findings: We conducted a genome-wide association study (GWAS) in 1,213 SCA (HbSS/HbSβ0) patients in Tanzania. Genotyping was done with Illumina Omni2.5 array and imputation using 1000 Genomes Phase I release data. Association with HbF was analysed using a linear mixed model to control for complex population structure within our study. We successfully replicated known associations for HbF near BCL11A and the HBS1L-MYB intergenic polymorphisms (HMIP), including multiple independent effects near BCL11A, consistent with previous reports. We observed eight additional associations with P<10(-6). These associations could not be replicated in a SCA population in the UK.

Conclusions: This is the largest GWAS study in SCA in Africa. We have confirmed known associations and identified new genetic associations with HbF that require further replication in SCA populations in Africa.

Show MeSH
Related in: MedlinePlus