Combining Evidence of Natural Selection with Association Analysis Increases Power to Detect Malaria-Resistance Variants
2007; Elsevier BV; Volume: 81; Issue: 2 Linguagem: Inglês
10.1086/519221
ISSN1537-6605
AutoresGeorge Ayodo, Alkes L. Price, Alon Keinan, Arthur Ajwang, Michael F. Otieno, Alloys S. S. Orago, Nick Patterson, David Reich,
Tópico(s)HIV Research and Treatment
ResumoStatistical power to detect disease variants can be increased by weighting candidates by their evidence of natural selection. To demonstrate that this theoretical idea works in practice, we performed an association study of 10 putative resistance variants in 471 severe malaria cases and 474 controls from the Luo in Kenya. We replicated associations at HBB (P=.0008) and CD36 (P=.03) but also showed that the same variants are unusually differentiated in frequency between the Luo and Yoruba (who historically have been exposed to malaria) and the Masai and Kikuyu (who have not been exposed). This empirically demonstrates that combining association analysis with evidence of natural selection can increase power to detect risk variants by orders of magnitude—up to P=.000018 for HBB and P=.00043 for CD36. Statistical power to detect disease variants can be increased by weighting candidates by their evidence of natural selection. To demonstrate that this theoretical idea works in practice, we performed an association study of 10 putative resistance variants in 471 severe malaria cases and 474 controls from the Luo in Kenya. We replicated associations at HBB (P=.0008) and CD36 (P=.03) but also showed that the same variants are unusually differentiated in frequency between the Luo and Yoruba (who historically have been exposed to malaria) and the Masai and Kikuyu (who have not been exposed). This empirically demonstrates that combining association analysis with evidence of natural selection can increase power to detect risk variants by orders of magnitude—up to P=.000018 for HBB and P=.00043 for CD36. Malaria infection (MIM 248310) has exerted severe pressure on the human genome within the past 10,000 years,1Bamshad M Wooding SP Signatures of natural selection in the human genome.Nat Rev Genet. 2003; 4: 99-111Crossref PubMed Scopus (330) Google Scholar, 2Tishkoff SA Varkonyi R Cahinhinan N Abbes S Argyropoulos G Destro-Bisol G Drousiotou A Dangerfield B Lefranc G Loiselet J et al.Haplotype diversity and linkage disequilibrium at human G6PD: recent origin of alleles that confer malarial resistance.Science. 2001; 293: 455-462Crossref PubMed Scopus (446) Google Scholar, 3Sabeti PC Reich DE Higgins JM Levine HZ Richter DJ Schaffner SF Gabriel SB Platko JV Patterson NJ McDonald GJ et al.Detecting recent positive selection in the human genome from haplotype structure.Nature. 2002; 419: 832-837Crossref PubMed Scopus (1319) Google Scholar and there are more cases today than ever before, with an estimated 300–660 million new episodes of clinical Plasmodium falciparum malaria every year.4Guerra CA Snow RW Hay SI Defining the global spatial limits of malaria transmission in 2005.Adv Parasitol. 2006; 62: 157-179Crossref PubMed Scopus (54) Google Scholar Despite high infection rates, only 1%–2% of patients develop life-threatening complications, such as cerebral malaria and profound anemia,5Kwiatkowski DP How malaria has affected the human genome and what human genetics can teach us about malaria.Am J Hum Genet. 2005; 77: 171-192Abstract Full Text Full Text PDF PubMed Scopus (677) Google Scholar so natural selection has likely operated, to a large extent, on severity. In the context of high infection rates, the genetics of host response are likely to play an important role.6Mackinnon MJ Mwangi TW Snow RW Marsh K Williams TN Heritability of malaria in Africa.PLoS Med. 2005; 2: e340Crossref PubMed Scopus (165) Google Scholar In sub-Saharan Africa, the populations in which malaria is endemic generally have a lower proportion of cases with severe disease.5Kwiatkowski DP How malaria has affected the human genome and what human genetics can teach us about malaria.Am J Hum Genet. 2005; 77: 171-192Abstract Full Text Full Text PDF PubMed Scopus (677) Google Scholar, 7Clarke SE Brooker S Njagi JK Njau E Estambale B Muchiri E Magnussen P Malaria morbidity among school children living in two areas of contrasting transmission in western Kenya.Am J Trop Med Hyg. 2004; 71: 732-738PubMed Google Scholar This suggests that there exist genetic variants that have risen to higher frequency in malaria-endemic populations because they modulate risk of P. falciparum malaria, similar to the case of the Duffy-null variant that protects against P. vivax malaria.8Miller LH Mason SJ Clyde DF McGinniss MH The resistance factor to Plasmodium vivax in blacks: the Duffy-blood-group genotype, FyFy.N Engl J Med. 1976; 295: 302-304Crossref PubMed Scopus (888) Google Scholar A handful of genetic variants have already been associated with risk of or protection against severe malaria infection.5Kwiatkowski DP How malaria has affected the human genome and what human genetics can teach us about malaria.Am J Hum Genet. 2005; 77: 171-192Abstract Full Text Full Text PDF PubMed Scopus (677) Google Scholar Our first objective in this study was to test variants of β-globin (HbAS9Modiano D Luoni G Sirima BS Simpore J Verra F Konate A Rastrelli E Olivieri A Calissano C Paganotti GM et al.Haemoglobin C protects against clinical Plasmodium falciparum malaria.Nature. 2001; 414: 305-308Crossref PubMed Scopus (255) Google Scholar, 10Aidoo M Terlouw DJ Kolczak MS McElroy PD ter Kuile FO Kariuki S Nahlen BL Lal AA Udhayakumar V Protective effects of the sickle cell gene against malaria morbidity and mortality.Lancet. 2002; 359: 1311-1312Abstract Full Text Full Text PDF PubMed Scopus (409) Google Scholar), intercellular adhesion molecule (ICAM TT11Kun JF Klabunde J Lell B Luckner D Alpers M May J Meyer C Kremsner PG Association of the ICAM-1Kilifi mutation with protection against severe malaria in Lambarene, Gabon.Am J Trop Med Hyg. 1999; 61: 776-779PubMed Google Scholar), CD36 (CD36 GT12Aitman TJ Cooper LD Norsworthy PJ Wahid FN Gray JK Curtis BR McKeigue PM Kwiatkowski D Greenwood BM Snow RW et al.Malaria susceptibility and CD36 mutation.Nature. 2000; 405: 1015-1016Crossref PubMed Scopus (178) Google Scholar), nitric oxide synthase (NOS2A 1659 AA13Burgner D Usen S Rockett K Jallow M Ackerman H Cervino A Pinder M Kwiatkowski DP Nucleotide and haplotypic diversity of the NOS2A promoter region and its relationship to cerebral malaria.Hum Genet. 2003; 112: 379-386PubMed Google Scholar), tumor necrosis factor (TNF 238 A14Knight JC Udalova I Hill AV Greenwood BM Peshu N Marsh K Kwiatkowski D A polymorphism that affects OCT-1 binding to the TNF promoter region is associated with severe malaria.Nat Genet. 1999; 22: 145-150Crossref PubMed Scopus (422) Google Scholar and TNF 308 A14Knight JC Udalova I Hill AV Greenwood BM Peshu N Marsh K Kwiatkowski D A polymorphism that affects OCT-1 binding to the TNF promoter region is associated with severe malaria.Nat Genet. 1999; 22: 145-150Crossref PubMed Scopus (422) Google Scholar, 15McGuire W Hill AV Allsopp CE Greenwood BM Kwiatkowski D Variation in the TNF-alpha promoter region associated with susceptibility to cerebral malaria.Nature. 1994; 371: 508-510Crossref PubMed Scopus (1086) Google Scholar, 16Flori L Delahaye NF Iraqi FA Hernandez-Valladares M Fumoux F Rihet P TNF as a malaria candidate gene: polymorphism-screening and family-based association analysis of mild malaria attack and parasitemia in Burkina Faso.Genes Immun. 2005; 6: 472-480Crossref PubMed Scopus (47) Google Scholar), Fc γ-receptor IIA (CD32 AA17Shi YP Nahlen BL Kariuki S Urdahl KB McElroy PD Roberts JM Lal AA Fcγ receptor IIa (CD32) polymorphism is associated with protection of infants against high-density Plasmodium falciparum infection. VII. Asembo Bay Cohort Project.J Infect Dis. 2001; 184: 107-111Crossref PubMed Scopus (68) Google Scholar, 18Cooke GS Aucan C Walley AJ Segal S Greenwood BM Kwiatkowski DP Hill AV Association of Fcgamma receptor IIa (CD32) polymorphism with severe malaria in West Africa.Am J Trop Med Hyg. 2003; 69: 565-568PubMed Google Scholar), interferon-α receptor-1 (IFNARI LI168V CC19Aucan C Walley AJ Hennig BJ Fitness J Frodsham A Zhang L Kwiatkowski D Hill AV Interferon-alpha receptor-1 (IFNAR1) variants are associated with protection against cerebral malaria in the Gambia.Genes Immun. 2003; 4: 275-282Crossref PubMed Scopus (72) Google Scholar and IFNARI 17470 CC19Aucan C Walley AJ Hennig BJ Fitness J Frodsham A Zhang L Kwiatkowski D Hill AV Interferon-alpha receptor-1 (IFNAR1) variants are associated with protection against cerebral malaria in the Gambia.Genes Immun. 2003; 4: 275-282Crossref PubMed Scopus (72) Google Scholar), and Toll-like receptor (TLR420Mockenhaupt FP Cramer JP Hamann L Stegemann MS Eckert J Oh NR Otchwemah RN Dietz E Ehrhardt S Schroder NW et al.Toll-like receptor (TLR) polymorphisms in African children: common TLR-4 variants predispose to severe malaria.Proc Natl Acad Sci USA. 2006; 103: 177-182Crossref PubMed Scopus (215) Google Scholar), which had previously been associated with malaria susceptibility. The particular phenotype we focused on was high levels of parasitemia in young children due to malaria infection. Second, we compared the frequency differentiation in populations in which malaria is endemic and in closely related populations in which it is not endemic, searching for the differences that would be expected if natural selection had affected those alleles in one population but not in the other, because malaria began to affect only one group. Finally, we formally combined the evidence of association from case-control studies with evidence of natural selection in populations that have been exposed to malaria infection. We note that there has been discussion elsewhere of how one could formally combine case-control association studies with statistical weights obtained on the basis of evidence of natural selection.21Roeder K Bacanu SA Wasserman L Devlin B Using linkage genome scans to improve power of association in genome scans.Am J Hum Genet. 2006; 78: 243-252Abstract Full Text Full Text PDF PubMed Scopus (159) Google Scholar Our goal in this study was to empirically demonstrate the power of this approach. We collected 471 severe malaria cases and 474 controls from the Luo ethnic group, a population that speaks a Nilotic language and lives in a malaria-endemic region in western Kenya. All the severe malaria cases were collected from the Bondo District Hospital's children's emergency ward or from its outpatient clinic between May 2004 and August 2005. The average age of the cases was 2.6 years (table 1), reflecting our focus on individuals with no previous immunological protection against malaria. The controls were randomly collected from volunteers at nearby secondary schools, with an average age of 16.9 years (table 1). We focused on older controls, because we knew that they had survived to an older age. Thus, the control samples selected for this study may be slightly enriched for variants protecting against severe malaria, which should make it slightly easier to detect associations.Table 1Characteristics of the Populations Included in This StudyPopulationMean Age (Range)No. in Sample (Male/Female)SourceMalaria EndemicityAltitude above Sea Level (m)Luo cases2.6 (1.5–10.0)471 (232/239)Bondo District Hospital, KenyaEndemic∼1,240Luo controls16.9 (14–20)474 (290/184)Bondo schools, KenyaEndemic∼1,240Masai controls16.9 (13–21)97 (42/55)Narok schools, KenyaNonendemic∼1,880Kikuyu controls17.1 (15–19)110 (46/64)Nyeri schools, KenyaNonendemic∼1,950Yoruba controlsNAaNA = not available.55 (27/28)International Haplotype MapEndemic∼700a NA = not available. Open table in a new tab For the selection study, we assembled population control samples from the Masai, Kikuyu, and Yoruba ethnic groups. We collected samples from the Masai and Kikuyu from secondary schools in Narok and Nyeri, Kenya, respectively (table 1). The Yoruba samples were from the International Haplotype Map project22The International HapMap Consortium A haplotype map of the human genome.Nature. 2005; 437: 1299-1320Crossref PubMed Scopus (4547) Google Scholar; we analyzed data from unrelated men and women, the parents in HapMap mother-father-child trios. About 2 ml of blood was obtained by venipuncture for all the samples we collected in Kenya. We extracted DNA within 10 h of blood collection, using a Qiagen DNA Blood mini kit, and then stored it at −20°C. All the participants provided informed consent, and, for children, informed consent was obtained from the parents and/or guardians. The study was reviewed and approved by the Harvard Medical School and Kenyatta University ethical review boards and by the Kenyan government. We identified human subjects who had severe malaria according to World Health Organization criteria. Blood smears and Giemsa staining were used to determine the asexual parasite count (parasitemia level). We identified cases as young children with >12 parasites per 200 red blood cells. All cases were also required to have overlapping clinical manifestations at the time of hospitalization, such as respiratory distress, convulsions, prostration, and hyperthermia (>39°C). We genotyped all human subjects for 13 candidate malaria SNPs, using mass spectrometry (Sequenom).23Tang K Fu DJ Julien D Braun A Cantor CR Koster H Chip-based genotyping by mass spectrometry.Proc Natl Acad Sci USA. 1999; 96: 10016-10020Crossref PubMed Scopus (230) Google Scholar We discarded SNPs with minor-allele frequency averaging <5% across the four ethnic groups, leaving 10 SNPs for subsequent analysis (table 2). Although the X-linked G6PD and CD40 genes are important candidates for malaria-resistance genes,3Sabeti PC Reich DE Higgins JM Levine HZ Richter DJ Schaffner SF Gabriel SB Platko JV Patterson NJ McDonald GJ et al.Detecting recent positive selection in the human genome from haplotype structure.Nature. 2002; 419: 832-837Crossref PubMed Scopus (1319) Google Scholar we excluded them from this study because we wished to focus on autosomal SNPs that we could compare with an empirical panel of autosomal variants in the genome.Table 2Replication Analysis for 10 Genotypes or Alleles Previously Associated with Malaria SusceptibilityGenotype or AlleleReference SNPDirection of Previous AssociationFrequency in Controls (%)No. of Cases/Controls GenotypedOR (95% CI)PHbAS10Aidoo M Terlouw DJ Kolczak MS McElroy PD ter Kuile FO Kariuki S Nahlen BL Lal AA Udhayakumar V Protective effects of the sickle cell gene against malaria morbidity and mortality.Lancet. 2002; 359: 1311-1312Abstract Full Text Full Text PDF PubMed Scopus (409) Google Scholarrs334ProtectionaPreviously published association with severe malaria.,bPreviously published association with mild malaria.25447/454.57 (.41–.79).0004CD36 GT12Aitman TJ Cooper LD Norsworthy PJ Wahid FN Gray JK Curtis BR McKeigue PM Kwiatkowski D Greenwood BM Snow RW et al.Malaria susceptibility and CD36 mutation.Nature. 2000; 405: 1015-1016Crossref PubMed Scopus (178) Google Scholarrs3211938RiskcPreviously published association with cerebral malaria.12456/4571.50 (1.03–2.18).015ICAM TT11Kun JF Klabunde J Lell B Luckner D Alpers M May J Meyer C Kremsner PG Association of the ICAM-1Kilifi mutation with protection against severe malaria in Lambarene, Gabon.Am J Trop Med Hyg. 1999; 61: 776-779PubMed Google Scholarrs5491ProtectionaPreviously published association with severe malaria.7460/455.71 (.42–1.21).10NOS2A 1659 AA13Burgner D Usen S Rockett K Jallow M Ackerman H Cervino A Pinder M Kwiatkowski DP Nucleotide and haplotypic diversity of the NOS2A promoter region and its relationship to cerebral malaria.Hum Genet. 2003; 112: 379-386PubMed Google Scholarrs8078340RiskcPreviously published association with cerebral malaria.,dPreviously published association with severe malarial anemia.6450/455.42 (.21–.83).99TNF 238 A14Knight JC Udalova I Hill AV Greenwood BM Peshu N Marsh K Kwiatkowski D A polymorphism that affects OCT-1 binding to the TNF promoter region is associated with severe malaria.Nat Genet. 1999; 22: 145-150Crossref PubMed Scopus (422) Google Scholar, 15McGuire W Hill AV Allsopp CE Greenwood BM Kwiatkowski D Variation in the TNF-alpha promoter region associated with susceptibility to cerebral malaria.Nature. 1994; 371: 508-510Crossref PubMed Scopus (1086) Google Scholarrs361525RiskcPreviously published association with cerebral malaria.9459/4571.00 (.73–1.39).49CD32 AA17Shi YP Nahlen BL Kariuki S Urdahl KB McElroy PD Roberts JM Lal AA Fcγ receptor IIa (CD32) polymorphism is associated with protection of infants against high-density Plasmodium falciparum infection. VII. Asembo Bay Cohort Project.J Infect Dis. 2001; 184: 107-111Crossref PubMed Scopus (68) Google Scholar, 18Cooke GS Aucan C Walley AJ Segal S Greenwood BM Kwiatkowski DP Hill AV Association of Fcgamma receptor IIa (CD32) polymorphism with severe malaria in West Africa.Am J Trop Med Hyg. 2003; 69: 565-568PubMed Google Scholarrs1801274ProtectiondPreviously published association with severe malarial anemia.,ePreviously published association with parasitemia.25455/447.95 (.71–1.29).38IFNARI LI168V CC19Aucan C Walley AJ Hennig BJ Fitness J Frodsham A Zhang L Kwiatkowski D Hill AV Interferon-alpha receptor-1 (IFNAR1) variants are associated with protection against cerebral malaria in the Gambia.Genes Immun. 2003; 4: 275-282Crossref PubMed Scopus (72) Google Scholarrs2257167ProtectioncPreviously published association with cerebral malaria.3455/4571.18 (.54–2.07).76TNF 308 A14Knight JC Udalova I Hill AV Greenwood BM Peshu N Marsh K Kwiatkowski D A polymorphism that affects OCT-1 binding to the TNF promoter region is associated with severe malaria.Nat Genet. 1999; 22: 145-150Crossref PubMed Scopus (422) Google Scholar, 16Flori L Delahaye NF Iraqi FA Hernandez-Valladares M Fumoux F Rihet P TNF as a malaria candidate gene: polymorphism-screening and family-based association analysis of mild malaria attack and parasitemia in Burkina Faso.Genes Immun. 2005; 6: 472-480Crossref PubMed Scopus (47) Google Scholarrs1800629RiskcPreviously published association with cerebral malaria.9450/4331.13 (.82–1.56).21IFNARI 17470 CC19Aucan C Walley AJ Hennig BJ Fitness J Frodsham A Zhang L Kwiatkowski D Hill AV Interferon-alpha receptor-1 (IFNAR1) variants are associated with protection against cerebral malaria in the Gambia.Genes Immun. 2003; 4: 275-282Crossref PubMed Scopus (72) Google Scholarrs1012335ProtectioncPreviously published association with cerebral malaria.3455/452.85 (.53–1.36).34TLR4 AG20Mockenhaupt FP Cramer JP Hamann L Stegemann MS Eckert J Oh NR Otchwemah RN Dietz E Ehrhardt S Schroder NW et al.Toll-like receptor (TLR) polymorphisms in African children: common TLR-4 variants predispose to severe malaria.Proc Natl Acad Sci USA. 2006; 103: 177-182Crossref PubMed Scopus (215) Google Scholarrs4986790RiskaPreviously published association with severe malaria.10407/3031.36 (.85–2.17).10a Previously published association with severe malaria.b Previously published association with mild malaria.c Previously published association with cerebral malaria.d Previously published association with severe malarial anemia.e Previously published association with parasitemia. Open table in a new tab As an assessment of genotyping quality, we observed that, for 85 genotypes obtained in duplicate, there were 2 discrepancies, for a discordance rate of 2.4%. After removing samples with .05). For the assessment of allele-frequency differentiation at random SNPs, we used the Illumina Bead Lab System to genotype 1,536 random SNPs from the Illumina linkage panel (covering chromosomes 1, 2, 3, and 22) in 45 of the Luo controls, 47 Masai controls, and 37 Kikuyu controls. We also obtained genotypes for these SNPs in 55 Yoruba samples from the HapMap database.22The International HapMap Consortium A haplotype map of the human genome.Nature. 2005; 437: 1299-1320Crossref PubMed Scopus (4547) Google Scholar Of these SNPs, 1,454 passed standard quality checks and had been genotyped in all four populations. We assessed the statistical significance of allele-frequency differences between Luo cases and Luo controls, using a χ2 test with 1 df. We used a one-tailed test of statistical significance, since our interest was in assessing whether a genotype or allele previously associated with malaria is more common in cases than in controls. We computed odds ratios (ORs) as A=(fcase/1-fcase)/(fcontrol/1-fcontrol), where fcase is the frequency in cases and fcontrol is the frequency in controls. We also computed a 95% CI as the range of ORs that produced a likelihood ratio consistent with the data (P>.05). Specifically, we estimated the SE of the log OR asB=(1ncase-ref+1ncase-var+1ncontrol-ref+1ncontrol-var)0.5, where ncase-ref and ncase-var are the counts of the reference and variant genotypes in cases, and ncontrol-ref and ncontrol-var are the analagous quantities in controls. The 95% CI is quoted as the range (eln(A)-1.65B to eln(A)+1.65B). To test for possible epistasis between any two SNPs, we used logistic regression. We compared the fit of three models with the data (case-control status for all the Luo samples): (1) genotype at the first SNP, (2) genotype at the second SNP, and (3) genotype at both SNPs.24Hosmer DW Lemeshow S Applied logistic regression. Wiley, New York1989Google Scholar We performed a one-tailed test for association with the genotypes previously associated with malaria. We calculated a Wald statistic and assessed significance for the epistatic interaction by a χ2 test with 1 df. The model of allele-frequency differentiation between two populations that we used to test for selection is that the difference in population frequencies at a given polymorphism is normally distributed with mean 0 and variance cp(1−p), where p is the ancestral frequency. This model is similar to that of Nicholson et al.,25Nicholson G Smith AV Jonsson F Gustafsson O Stefansson K Donnelly P Assessing population differentiation and isolation from single nucleotide polymorphism data.J R Stat Soc. 2002; 64: 695-715Crossref Scopus (158) Google Scholar who showed that, for populations with modest genetic divergence times, it is a good approximation for allele-frequency differentiation. Under certain assumptions, the c parameter is expected to equal 2×FST. From a population genetics perspective, c can be viewed as measuring genetic drift between populations. To estimate c empirically, we used data from the 1,454 randomly chosen markers. For a given pair of populations, we estimated c as the empirical variance of the difference in population frequencies, after normalizing by p(1−p) and accounting for sampling noise, which has variance p(1-p)(1/N1+1/N2), where N1 and N2 are total allele counts for the two populations at a given marker. We approximated the normalization term p(1−p) by setting p equal to the average of observed frequencies of the two populations, and we approximated binomial sampling noise as normally distributed. The same approximations were applied both to our estimation of c and to our subsequent analysis of individual markers. SNPs with average minor-allele frequency <5% for the two populations being compared were omitted from all computations, since the normal approximation becomes less reliable (table 3).Table 3Tests for Differentiating Selection between Malaria-Endemic and -Nonendemic PopulationsAllele Frequency (%) (No. of Alleles Used in Assessment)PaP values for selection are based on allele-frequency differentiation tests between malaria-endemic (Luo and Yoruba) and -nonendemic populations (Kikuyu and Masai). Values in bold are significant. Statistics for SNPs with average minor-allele frequency <5% for the two populations analyzed are denoted as NA (not available).AlleleReference SNPLuoYorubaMasaiKikuyuLuo vs. MasaiLuo vs. KikuyuYoruba vs. MasaiYoruba vs. KikuyuHbAS Trs33413 (908)11 (102)0 (194)0 (200).00149.00036.044.025CD36 Grs32119386 (914)22 (100)1 (186)0 (202)NANA.00590.00096ICAM Trs549125 (910)24 (100)16 (186)18 (206).19.25.41.48NOS2A 1659 Ars807834021 (910)19 (98)25 (188)21 (204).62.86.57.89TNF 238 Ars3615259 (914)1 (100)21 (192)16 (202).04.13.00741.010CD32 Ars180127450 (900)50 (98)50 (190)44 (210)1.0.371.0.56IFNARI L168V Crs225716716 (914)16 (98)25 (192)19 (208).20.61.37.72TNF 308 Ars18006299 (866)6 (96)6 (188)7 (204).60.741.0.84IFNARI 17470 Crs101233532 (904)22 (108)33 (190)35 (202).92.64.32.18TLR4 Grs49867905 (633)4 (114)7 (178)5 (170).601.0.6.84a P values for selection are based on allele-frequency differentiation tests between malaria-endemic (Luo and Yoruba) and -nonendemic populations (Kikuyu and Masai). Values in bold are significant. Statistics for SNPs with average minor-allele frequency <5% for the two populations analyzed are denoted as NA (not available). Open table in a new tab To test whether an individual marker was more differentiated than expected between two populations, we compared the observed difference in frequency with the expected distribution N[0,p(1-i)(c+1/N1+1/N2)], using the value of c estimated above, and computed a χ2 statistic with 1 df. A feature of this test is that the χ2 statistic has a mean value of 1 across the set of markers used to infer c. The test appropriately handles different sample sizes for candidate markers versus random markers used to infer c. A detailed statistical treatment will appear elsewhere (A.L.P, N.P., and D.R., unpublished data). The combined test formally evaluates whether the observed data are consistent with the model of no case-control association and no selection. The test is performed by summing the association χ2 statistic and the differentiation χ2 statistic, forming a χ2 statistic with 2 df. We note that the association χ2 statistic used in this test is, by definition, a two-tailed statistic. We computed this sum for each pair of populations, using the same association statistic in each case. When one of the two populations being compared was the Luo population, we used the summed counts of Luo cases and Luo controls in the combined statistics reported in table 4. This generally leads to less significant P values than does using Luo controls only (and so is conservative). Using summed counts of Luo cases and Luo controls is appropriate under the null assumption of no association and ensures that the association statistic and differentiation statistic are independent. However, for the selection-only statistics reported in table 3, we used Luo controls only, since we wished to evaluate the evidence of selection in the control population, without regard to evidence of case-control association.Table 4Formal Combination of Case-Control Association Analysis and Tests of Natural SelectionPaP values from combining case-control association studies with the test for differentiating selection between malaria-endemic (Luo and Yoruba) and -nonendemic (Masai and Kikuyu) populations. Values in bold are significant. NA = not available.AlleleReference SNPLuo vs. MasaiLuo vs. KikuyuYoruba vs. MasaiYoruba vs. KikuyuHbAS Trs334.000056.000018.00048.00029CD36 Grs3211938NANA.0023.00043ICAM Trs5491.19.23.32.36NOS2A 1659 Ars8078340.033.038.032.038TNF 238 Ars361525.12.30.028.038CD32 Ars1801274.96.65.96.81IFNARI L168V Crs2257167.34.68.52.73TNF 308 Ars1800629.63.69.76.75IFNARI 17470 Crs1012335.91.79.56.38TLR4 Grs4986790.42.41.38.42a P values from combining case-control association studies with the test for differentiating selection between malaria-endemic (Luo and Yoruba) and -nonendemic (Masai and Kikuyu) populations. Values in bold are significant. NA = not available. Open table in a new tab We tested each of the 10 variants for association with malaria, comparing Luo cases with Luo controls. Two of the variants showed nominally statistically significant associations by one-tailed tests that searched for an association with the genotype or allele previously proposed to affect malaria resistance (table 3). We replicated the well-known association in which heterozygotes for the sickle-cell trait HbAS (HbAS T) are protected against severe malaria (P=.0004; OR 0.57 [95% CI 0.41–0.79]) (see the "Material and Methods" section). Although the OR of 0.57 is less strong than that observed in some previous studies,9Modiano D Luoni G Sirima BS Simpore J Verra F Konate A Rastrelli E Olivieri A Calissano C Paganotti GM et al.Haemoglobin C protects against clinical Plasmodium falciparum malaria.Nature. 2001; 414: 305-308Crossref PubMed Scopus (255) Google Scholar it is in the same range as the OR of 0.45 (0.24–0.84), which was observed in another study of young children with a similar phenotype of severe malaria.10Aidoo M Terlouw DJ Kolczak MS McElroy PD ter Kuile FO Kariuki S Nahlen BL Lal AA Udhayakumar V Protective effects of the sickle cell gene against malaria morbidity and mortality.Lancet. 2002; 359: 1311-1312Abstract Full Text Full Text PDF PubMed Scopus (409) Google Scholar Different case-control studies focus on different phenotypes, and the protection of HbAS against severe malaria is known to vary with age,26Williams TN Mwangi TW Roberts DJ Alexander ND Weatherall DJ Wambua S Kortok M Snow RW Marsh K An immune basis for malaria protection by the sickle cell trait.PLoS Med. 2005; 2: e128Crossref PubMed Scopus (139) Google Scholar so it is not surprising that the estimated ORs are heterogeneous across studies. We also replicated the association in which heterozygotes for CD36 GT are at increased risk for se
Referência(s)