MICA polymorphism in South American Indians

We have studied the MICA alleles of 196 unrelated subjects from three South American Indian tribes (Toba, Wichi and Terena). They are members of isolated tribes located in the Gran Chaco area in northeastern Argentina and in Mato Grosso do Sul in South Central Brazil. Of 55 previously known alleles, nine were observed in South American Indians, compared with 16 that were found in North American Caucasians, suggesting a more restricted allelic distribution of MICA in these tribes. In South American Indians, MICA*00201 was the most frequent allele, with a gene frequency of 33% in Toba, 47% in Wichi and 44% in Terena. MICA*00201, MICA*027 (external domain sequence like MICA*008/TM allele A5) and MICA*010 accounted for more than 90% of all the MICA genes in South American Indians. In North American Caucasians, MICA*00801 (*008/A5.1) accounted for 42% of the genes and was the most common allele. We observed a high degree of linkage disequilibrium between certain alleles of MICA and of HLA-B in the South American Indian populations. Phylogenetic trees constructed using gene frequencies of the transmembrane short tandem repeats in the populations reported here, and in other populations taken from published reports, suggest that South American Indians are more closely related to Asians than to Europeans.     

Possible influence of BCHE locus of butyrylcholinesterase on stature and body mass index

Butyrylcholinesterase activity has been shown to be positively associated with weight and body mass index (BMI). The present study was carried out to search for an association between variants of the BCHE gene and weight, stature, and BMI on the basis of means and variances compared between nonusual variants and their respective usual controls. Individuals bearing the atypical mutation (N = 52) did not differ from their usual phenotype controls (N = 104) in these parameters. The BCHE*U/BCHE*K individuals (N = 222) presented a significantly higher BMI variance than their BCHE*U/BCHE*U controls (N = 222, F = 1.40, P = 0.012). This higher BMI variance does not seem to be an isolated effect of the K mutation, but appears to be the result of an interaction between the K allele and the usual allele, since no such difference in variance was detected between BCHE*K/BCHE*K individuals (N = 23) and their BCHE*U/BCHE*U (N = 23) controls. These data may suggest a relation between variability in the BCHE locus itself and BMI. Individuals with the BCHE UF phenotype (N = 45) showed a significantly higher mean stature (about 3 cm more; P = 0.02) than their controls with the usual phenotype (N = 135). A role in cell proliferation has been proposed for BCHE, and since growth depends on the number of mitoses, it is not unexpected that variants of this enzyme may influence body stature in different ways. This study reports the first data on the relation of BCHE alleles to anthropometric characters.     

Electrophoretic and isoelectric focusing studies in Brazilian Indians: data on four systems

Three-hundred ninety-nine individuals living in seven populations of two Brazilian Indian tribes (Macushi and I?ana River Indians) were tested for the phosphoglucomutase 1 (PGM1), properdin factor B (BF), haptoglobin (HP), and alpha-1-antitrypsin (PI) systems. We observed significant internal heterogeneity in the two tribes for the PGM1 alleles and in the Macushi for the HP markers. Frequencies in three of the four systems (the exception being BF) also show clear differences in the Macushi and I?ana River Indians. Compared with other ethnic groups, South American Indians generally present high frequencies of PGM1*1B, BF*S, HP*1S, and PI*M3. On the other hand, PGM1*1A, PI*M1, and PI*M2 are reduced, and HP*1F is absent or rare. This is the first report about HP subtypes among American Indians.     

Apolipoprotein E genotyping among the healthy Kuwaiti population

Apolipoproteins (lipid-free) are lipid-binding proteins that circulate in the plasma of human blood and are responsible for the clearance of lipoproteins. Apolipoprotein E (ApoE) is one of the several classes of this protein family. It acts as a ligand for the low-density lipid (LDL) receptors and is important for the clearance of very low-density lipid (VLDL) and chylomicron remnants. The APOE gene locus is polymorphic, with three major known alleles, APOE*3, *4, and *2. We investigated the distribution of the allele frequency of the APOE gene locus and describe here the genetic variation in four Kuwaiti subpopulations: Arab origin (Arabian peninsula), Arab Bedouin tribes, Iranian origin, and the heterogeneous population. We also describe the use of Spreadex gels in resolving the amplified and digested products of the APOE gene locus. DNA was extracted from whole blood and subjected to PCR and then to RFLP analysis. Allele and genotype frequencies were estimated for the total population and for each subpopulation. Statistical analysis showed no difference in the allele frequencies between the four groups. The frequency of APOE*3 in the Kuwaiti population was highest (88.4%) followed by the frequency of APOE*4 (6.5%) and APOE*2 (5.1%). The genotype and allele frequencies obtained for the Kuwaiti population fell within the reported worldwide distribution for the APOE gene locus. Moreover, the results obtained in this study showed no statistical difference (p > 0.05) between the APOE allele and genotype frequencies between the subgroups for all six genotypes and three alleles, supporting the assumption of admixture in the Kuwaiti population and that the obtained frequencies were in Hardy-Weinberg equilibrium. Finally, we found that the distribution of the APOE alleles in Kuwait differs somewhat from those reported in other Arab populations, suggesting that the Arabs originating from the Arabian peninsula are different from those of Lebanon, Morocco, and Sudan.     

HLA antigens and other genetic markers in the Mapuche Indians of Argentina

A total of 107 Mapuche Indians living in western Argentina were studied with respect to 16 genetic systems. For HLA, there were a few differences in relation to previous studies; and considering the averages observed in 15 other South American tribes, Mapuche Indians showed low values for A2, A9 and C3, but high ones for A28 and B16. This is the first report of the presence (in low frequencies, 1-6%) of alleles C2, C6 and C7, as well as of DR antigens (most frequent alleles DR4 and DR2) in South American Indians. Some peculiar reactions shown by products of locus B suggest the presence of antigens that are characteristic of the Mapuche. As for the other systems, the frequencies of R1 (Rh) and PGM1(1) were lower but those for r (Rh), GLO1 and Hp1 were higher than the averages obtained considering previous studies of this ethnic group. Other salient findings were the variability observed in the PGM2 and C3 systems, and the low prevalence of Bfs.     

Genotype frequencies of polymorphic GSTM1, GSTT1, and cytochrome P450 CYP1A1 in Mexicans

The genotype frequencies of three metabolic polymorphisms were determined in a sample of a typical community in central Mexico. CYP1A1*3, GSTM1, and GSTT1 polymorphisms were studied in 150 donors born in Mexico and with Mexican ascendants; with respect to ethnicity the subjects can be considered Mestizos. PCR reactions were used to amplify specific fragments of the selected genes from genomic DNA. An unexpected 56.7% frequency of the CYP1A1*3 allele (which depends on the presence of a Val residue in the 462 position of the enzyme, instead of Ile) was found, the highest described for open populations of different ethnic origins (i.e., Caucasian, Asian, African, or African American). The GSTM1 null genotype was found with a frequency of 42.6%, which is not different from other ethnicities, whereas the GSTT1 null genotype had a frequency of 9.3%, one of the lowest described for any ethnic group but comparable to the frequency found in India (9.7%). The frequency of the combined genotype CYP1A1*3/*3 and the GSTM1 null allele is one of the highest observed to date (or perhaps the highest): 13.7% among all the ethnicities studied, including Caucasians and Asians, whereas the combination of CYP1A1*3/*3 with the GSTT1 null allele reached only 2.8%. The GSTM1 null allele combined with the GSTT1 null allele, on the other hand, has one of the lowest frequencies described, 4.24%, comparable to the frequencies found in African Americans and Indians. Finally, the combined CYP1A1*3/*3, GSTM1 null allele, and GSTT1 null allele genotype could not be found in the sample studied; it is assumed that the frequency of carriers of these combined genotypes is less than 1%. CYP1A1*3 and CYP1A1*2 polymorphisms were also evaluated in 50 residents in a community of northern Mexico; the CYP1A1*3 frequency was 54%, similar to that found in the other community studied, and the CYP1A1*2 frequency was 40%, which is high compared to Caucasians and Asians but comparable to the frequency found in Japanese and lower than the frequency found in Mapuche Indians. Haplotype frequencies for these CYP1A1 polymorphisms were estimated, and a linkage disequilibrium value (D) of 0.137 was calculated.     

Distribution of hereditary blood groups among indians in South America. VII. In Argentina

This seventh and last paper in a series on the distribution of blood groups among Indians in South America reports the findings among Amerinds in Argentina. Blood specimens were procured from putative full-bloods of the following tribes: 38 Diaguita (Calchaqui), 230 Mataco, 90 Chiriguano, 142 Choroti, 51 Toba, 120 Chané, 96 Chulupi (Ashluslay), and 178 Araucano (Mapuche). These 945 samples were tested for blood factors in the A-B-O, M-N-S-s, P, Rh-Hr, K-k, Lewis, Duffy, Kidd, and Diego systems. Serum samples were tested for haptoglobins and transferrins. Hemolysates prepared from whole blood were tested for hemoglobin types. The results are presented in tables as phenotype distribution and calculated allele frequencies. Locations of the populations from which blood samples were procured are shown on a map of North and Central Argentina. High frequencies are reported for the O allele. Allele frequencies are high also for M, s, R¹ (CDe), R² (cDE), k, Le^H and Fy. They are usually low or absent for alleles B, N, S, Mi^a, Vw, R^o (cDe), r (cde), K, Le¹, and fy. The Di allele ranged from 0.013 in the Araucano (Mapuche) to 0.192 in the Toba. Allele frequencies aberrant for Indians were observed more often in the Araucano (Mapuche) and Diaguita tribes, due probably to greater inflow of non-Indian genes into their gene pool and perhaps also to genetic drift in small inbred populations. Hp¹ allele frequencies varied from 0.43 in the Choroti to 0.80 in the Diaguita. All samples tested for transferrins except six contained the variant Tf C; the six were B₁ C present in samples from one Mataco and six Araucano persons. All the specimens tested electrophoretically for hemoglobin types contained only (A) as a major component.     

Mexican American ancestry-informative markers: examination of population structure and marker characteristics in European Americans, Mexican Americans, Amerindians and Asians

Markers with large differences in allele frequencies between ethnicities provide ancestry information that can be applied to genetic studies. We identified over 100 biallelic ancestry informative markers (AIMs) with large allele frequency differences between European Americans (EA) and Pima Amerindians from laboratory and database screens. For 35 of these markers, Mayan, Yavapai and Quechuan Amerindians were genotyped and compared with EA and Pima allele frequencies. Markers with large allele frequency differences between EA and one Amerindian tribe showed only small differences between the Amerindian tribes. Examination of structure in individuals demonstrated a clear separation of subjects of European from those of Amerindian ancestry, and similarity between individuals from disparate Amerindian populations. The AIMs demonstrated the variation in ancestral composition of individual Mexican Americans, providing evidence of applicability in admixture mapping and in controlling for structure in association tests. In addition, a high percentage of single-nucleotide polymorphisms (SNPs) selected on the basis of large frequency differences between EA and Asian populations had large allele frequency differences between EA and Amerindians, suggesting an efficient method for greatly expanding AIMs for use in admixture mapping/structure analysis in Mexican Americans. Together, these data provide additional support for the practical application of admixture mapping in the Mexican American population.Electronic Supplementary Material Supplementary material is available in the online version of this article at     

Complement C3 polymorphism in sixteen populations of Colombia. Results of “Expedicion humana”

Blood samples of 1.022 individuals of both sexes, collected during the “Expedicion Humana” programme. Twelve Indian group, two Negroes, two urban and one mestizo populations of Colombia were C3 typed. The results showed the urban samples have C3*F genotype frequencies (17 and 20%) similar to the Spanish population. The Negroe system, with lower C3*F gene frequencies (3.5 and 7%) also resembled their parental populations in Africa. The Mestizo population reported in this study showed a lower value of the C3*F gene (2.2%) than previously studied. suggesting that the degree of admixture may vary in different regions of the country. As regards the Indian populations, there was a wide range of variation of the C3*F gene (0–19%) many tribes being monomorphic. This gene may be a marker of recent admixture in Amerindians, but there may be other genetic factors, such as selection and inbreeding in maintaining high levels of genetic polymorphism in tribes such as Guahibo.     

Characterization of 12 silent alleles of the human butyrylcholinesterase (BCHE) gene.

The silent phenotype of human butyrylcholinesterase (BChE), present in most human populations in frequencies of approximately 1/100,000, is characterized by the complete absence of BChE activity or by activity <10% of the average levels of the usual phenotype. Heterogeneity in this phenotype has been well established at the phenotypic level, but only a few silent BCHE alleles have been characterized at the DNA level. Twelve silent alleles of the human butyrylcholinesterase gene (BCHE) have been identified in 17 apparently unrelated patients who were selected by their increased sensitivity to the muscle relaxant succinylcholine. All of these alleles are characterized by single nucleotide substitutions or deletions leading to distinct changes in the structure of the BChE enzyme molecule. Nine of the nucleotide substitutions result in the replacement of single amino acid residues. Three of these variants, BCHE*33C, BCHE*198G, and BCHE*201T, produce normal amounts of immunoreactive but enzymatically inactive BChE protein in the plasma. The other six amino acid substitutions, encoded by BCHE*37S, BCHE*125F, BCHE*170E, BCHE*471R, and BCHE*518L, seem to cause reduced expression of BChE protein, and their role in determining the silent phenotype was confirmed by expression in cell culture. The other four silent alleles, BCHE*271STOP, BCHE*500STOP, BCHE*FS6, and BCHE*I2E3-8G, encode BChES truncated at their C-terminus because of premature stop codons caused by nucleotide substitutions, a frame shift, or altered splicing. The large number of different silent BCHE alleles found within a relatively small number of patients shows that the heterogeneity of the silent BChE phenotype is high. The characterization of silent BChE variants will be useful in the study of the structure/function relationship for this and other closely related enzymes.     

Allelic frequencies and statistical data obtained from 12 codis STR loci in an admixed population of the Brazilian Amazon

The allelic frequencies of 12 short tandem repeat loci were obtained from a sample of 307 unrelated individuals living in Macapá, a city in the northern Amazon region, Brazil. These loci are the most commonly used in forensics and paternity testing. Based on the allele frequency obtained for the population of Macapá, we estimated an interethnic admixture for the three parental groups (European, Native American and African) of, respectively, 46%, 35% and 19%. Comparing these allele frequencies with those of other Brazilian populations and of the Iberian Peninsula population, no significant distances were observed. The interpopulation genetic distances (F(ST) coefficients) to the present database ranged from F(ST) = 0.0016 between Macapá and Belém to F(ST) = 0.0036 between Macapá and the Iberian Peninsula.     

Further data on the microsatellite locus D12S67 in worldwide populations: an unusual distribution of D12S67 alleles in Native Americans

We report the frequencies of alleles at the microsatellite locus D12S67 in 2 widely separated ethnic groups of the world: 2 populations from Sulawesi, an island in the Indonesian archipelago, and 5 Native American tribes of Colombia, South America. The allele frequencies in the Minihasans and Torajans of Sulawesi are similar to each other (but the modal class allele is different) and in general agreement with those reported in mainland Asian groups, but different from both Europeans and Chinese Han of Taiwan. The 5 Native American tribes (Arsario, Kogui, Ijka, Wayuu, and Coreguaje) display different allele frequencies from those seen in Sulawesi populations, in other groups from Europe and mainland Asia, and in Chinese Han of Taiwan. Native Americans exhibit a bimodal distribution of alleles, unlike other groups, with significant differences among the tribes. The Arsario and Kogui have no admixture with Europeans or Africans and are the most distinctive, while the Wayuu have the most admixture and show most similarity to other groups. The data suggest that nonadmixed Native Americans may be quite distinctive with respect to this marker. The most common allele varies across the 5 tribes, from 249 base pairs to 261 base pairs. All samples exhibit Hardy-Weinberg genotype proportions; heterozygosities are lowest in the 2 nonadmixed Native American tribes. Examination of all the available data indicates that some east Asian and southeast Asian groups are characterized by a high frequency of smaller sized D12S67 alleles, while other populations have a greater proportion of the larger sized alleles. The cumulative, though still highly restricted, population data on locus D12S67 demonstrate that it may be of considerable value in anthropological genetic studies of ethnic groups. Data are required on Native Americans outside Colombia before this marker can be used in admixture studies of this group.     

Genetic variants of the paraoxonases (PON1 and PON2) in the Chilean population

We estimated the frequencies of PON1 and PON2 variants (linked genes) in two hospital samples taken from the northern (San José Hospital, SJH) and eastern (Clínica Las Condes, CLC) parts of Santiago, Chile, using the polymerase chain reaction followed by restriction endonuclease digestion. The two hospital samples have different degrees of Amerindian admixture (SJH, 34.5%; CLC, 15.9%), which is reflected in the observed frequencies of the PON1 *B allele (SJH, 43.1%; CLC, 33.7%) and the PON2*S allele (SJH, 86.3%; CLC, 77.6%); both allele frequencies are significantly different between samples. The frequencies of the combined PON1-PON2 genotypes *A/*B-*C/*C, *A/*B-*S/*S, and *B/*B-*S/*S and of the haplotypes PON*A,C and PON*B,S were significantly different between the SJH and CLC groups. None of the genotype frequencies deviated significantly from those predicted by the Hardy-Weinberg equation. No linkage disequilibrium was found between the PON1 alleles and any of the PON2 alleles in either group (all p > 0.05). In our samples 38.52% (SJH) and 26.25% (CLC) of chromosomes must have the haplotype PON*B,S, presumed to be related to the risk of coronary artery disease. Twenty-four of 193 (12.4%) SJH individuals and 7 of 122 (5.7%) CLC individuals were homozygotes for this haplotype. Finally, our data indicate ethnic-group-dependent genetic differences in the vulnerability to toxic organophosphorus.     

The mutation G298A-->Ala100Thr on the coding sequence of the Duffy antigen/chemokine receptor gene in non-caucasian Brazilians

Ala100Thr has been suggested to be a Caucasian genetic marker on the FY*B allele. As the Brazilian population has arisen from miscegenation among Portuguese, Africans, and Indians, this mutation could possibly be found in Euro- and Afro-Brazilians, or in Brazilian Indians. Fifty-three related individuals and a random sample of 100 subjects from the Brazilian population were investigated using the polymerase chain reaction and four restriction fragment length polymorphisms. Confirming the working hypothesis, among the related individuals three Afro-Brazilians (two of them a mother and daughter) and a woman of Amerindian descent had the Ala100Thr mutation on the FY*B allele. Five non-related Euro-Brazilians also carried the mutation. All nine individuals presented the Fy(a-b+) phenotype. We conclude that the Ala100Thr mutation can occur in populations other than Caucasians and that this mutation does not affect Duffy expression on red blood cells. Gene frequencies for this allele in the non-related individuals were in agreement with those of other populations. The Duffy frequencies of two Amerindian tribes were also investigated.     

Beta1 adrenergic receptor polymorphisms Arg389Gly and Ser49Gly in the Amerindian and Mestizo populations of Mexico

The beta1 adrenergic receptor genotypes (Ser49Gly and Arg389Gly) were determined in 190 individuals from 3 Mexican populations. Mestizos and Teenek present the highest frequencies for the *Arg allele and the lowest frequencies for the *Gly allele (Arg389Gly) compared to European, Asian, and African populations. Mayos present the highest frequency for the *Gly allele. The knowledge of the distribution of these alleles could help define the significance of these polymorphisms as genetic susceptibility markers in Amerindian populations.     

Interethnic and intraethnic variability of CYP2C8 and CYP2C9 polymorphisms in healthy individuals

Cytochrome P450 (CYP) superfamily members CYP2C8 and CYP2C9 are polymorphically expressed enzymes that are involved in the metabolic inactivation of several drugs, including, among others, antiepileptics, NSAIDs, oral hypoglycemics, and anticoagulants. Many of these drugs have a narrow therapeutic index, and growing evidence indicates a prominent role of CYP2C8 and CYP2C9 polymorphisms in the therapeutic efficacy and in the development of adverse effects among patients treated with drugs that are CYP2C8 or CYP2C9 substrates. In this review, we summarize present knowledge on human variability in the frequency of variant CYP2C8 and CYP2C9 alleles. Besides an expected interethnic variability in allele frequencies, a large intraethnic variability exists. Among Asian subjects, for example, statistically significant differences (p < 0.0001) in CYP2C9*3 allele frequencies between Chinese and Japanese individuals have been reported. In addition, individuals from East Asia present different allele frequencies for CYP2C9*2 and CYP2C9*3 compared with South Asian subjects (p < 0.0001). Among Caucasian Europeans, statistically significant differences for the frequency of CYP2C8*3, CYP2C9*2, and CYP2C9*3 exist (p < 0.0001). This indicates that Asian individuals or Caucasian European individuals cannot be considered as homogeneous groups regarding CYP2C8 or CYP2C9 allele frequencies. Caucasian American subjects also show a large variability in allele frequencies, which is likely to be related to ethnic ancestry. A higher frequency of variant CYP2C8 and CYP2C9 alleles is expected among Caucasian Americans with South European ancestry than in individuals with North European ancestry. The findings summarized in this review suggest that among individuals with Asian or European ancestry, intraethnic differences in the risk of developing adverse effects with drugs that are CYP2C8 or CYP2C9 substrates are to be expected. In addition, the observed intraethnic variability reinforces the need for proper selection of control subjects and points against the use of surrogate control groups for studies involving association of CYP2C8 or CYP2C9 alleles with adverse drug reactions or spontaneous diseases.     

Interaction between ABO and haptoglobin systems in a Bengalee population.

Blood samples from 2,232 individuals of a Bengalee Caste Hindu population were investigated in an attempt to confirm the association between the ABO and haptoglobin (HP) systems previously found in populations of European origin. Indians differ from Europeans in having lower HP*1 and higher ABO*B frequencies. In spite of this, as in previous studies, a weak HP/ABO association was found with a significantly lower HP*1 allele frequency in blood group O versus other ABO groups.     

Human MHC class III genes, Bf and C4. Polymorphism, complotypes and association with MHC class I genes in the Finnish population

Electrophoretically detected genetic polymorphism of human MHC class III genes, factor B (Bf) and complement C4A and C4B, was studied in the Finnish population. Bf alleles were determined in a panel of sera from 70 unrelated individuals. The common Bf alleles, Bf*S and Bf*F, had frequencies of 73% and 26%, respectively. Only in 1 individual was another allele, Bf*F1, detected. The frequencies of the C4A and C4B alleles were based on studies of 254 unrelated individuals. In this panel, five different alleles were detected at the C4A locus and four at the C4B locus. At both loci an allele without a gene product, i.e. a 'null' allele, was observed with high frequency, 11% for C4A 'null' and 17% for C4B 'null'. The association of complotypes to HLA haplotypes was analyzed in 70 chromosomes. The most common combination, defined by class I and class III alleles, was HLA-B7-S31 (13%), followed by HLA-B35-F20 (8.4%) and HLA-B8-S03 (7.1%). Some HLA-B specificities, for example B15, B27 and B40, were associated with a variety of complotypes. The importance of complotyping in HLA genetics is discussed.     

Haptoglobin 1F allele frequency is high among indigenous populations in the state of Durango, Mexico

We studied haptoglobin polymorphism in Mexican populations with high Indian ancestry living in isolated and urban areas in the state of Durango. Analysis with respect to the HP*1F and HP*1S allelic subtypes by isoelectric focusing showed unusually high HP*1F allele frequencies among urban (0.370) and isolated Mexican Indians (0.383). Comparison with other population studies demonstrated a geographical cline of the HP*1F allele increasing in the same direction of the HP*1 allele, while HP*1S frequency does not show racial differences.     

Apolipoprotein E and H polymorphisms in Mongolian Buryat: allele frequencies and relationship with plasma lipid levels

A Buryat population consisting of seven tribal groups in eastern Mongolia has been screened to determine the frequency distribution of different apolipoprotein E and H alleles (APOE and APOH, genes) coding for common isoforms and their association with quantitative plasma lipid levels. Allele frequencies at the APOE locus in 125 healthy Buryat aged 17 to 73 years were highest for APOE*3 (0.804), followed by APOE*4 (0.164) and APOE*2 (0.032). The APOH locus had high frequencies of APOH*2 (0.912) and APOH*3 (0.088). APOH*1 was not detected. No significant differences were observed in the overall APOE allele frequencies between the Buryat and the Siberian Evenki, Inuits, and Indians in Asia, or with some European whites. The frequency distribution of the overall APOH alleles of the Buryat was similar to that of the Japanese in Asia. Overall plasma lipid levels of the Buryat (males aged 20 to 73 years, females aged 21 to 64 years) were considerably lower, comparable to those of the Evenki. The APOE*4/E*3 males had significantly high total- and LDL-cholesterol levels compared with the APOE*3/E*3 males (p < 0.025 and p < 0.01, respectively). No significant effects of the APOH genotypes on any of the plasma lipid levels were observed. In particular, our data regarding APOE suggest that the Buryat are genetically close in allele frequencies to the Evenki and Inuits, but differ from them in the association of genotype APOE*4/E*3 with cholesterol levels.