Analysis of the Data from WTCCC
4.1.3 Test of association
Single SNP association
First of all, for the processed data above, we adopt Fisher’s exact test and Cochran-Armitage trend test for the genotypic test and the allele test respectively. Subjects are SNPs whose p-value is less than 5 × 10−7 (the strongest association, see table 4.1 and 4.2) or greater than 5 × 10−7 and less than 1 × 10−5 (moderate association, see table 4.3 and 4.4) for either the genotypic test or the allele test. Even though we found that the genetic variants evaluated the strongest and moderate associated with hypertension risk, some associated SNPs do not identify known genes or the relevance to hypertension.
Figure 4.2: Genome-wide Manhattan Plot for Hypertension on Single SNP-Based by Fisher’s Exact Test
CHRM2 (cholinergic receptor, muscarinic 2) belongs to a larger family of G protein-coupled receptors. The muscarinic cholinergic receptor 2 is involved in mediation of bradycardia and a decrease in cardiac contractility [Hautala et al., 2009]; [Zhang et al., 2008]. Carriers of the variant G of CHRM2 (rs7800093) has a significantly lower or higher risk of hypertension compared with individuals with the common homozygote genotype:
odds ratio [95% CI] for heterozygotes 0.02 [0.00-0.11] and for homozygotes 53.00 [12.98-216.38].
KCNB2 (potassium voltage-gated channel, Shab-related subfamily, member 2), the diverse functions of the protein include regulating neurotransmitter release, heart rate, insulin secretion, neuronal excitability, epithelial electrolyte transport, smooth muscle contraction, and cell volume. KCNB2 (rs11782342) has a significant increase in risk among homozygote variants: odds ratio [95% CI] = 1.98[1.56-2.53]. The association between KCNB2 and cardiovascular disease risk has been found in the previous study [Vasan et al., 2007].
HTR3B (5-hydroxytryptamine (serotonin) receptor 3B) encodes subunit B of the type 3 receptor for 5-hydroxytryptamine (serotonin), a biogenic hormone that functions as a neurotransmitter, a hormone, and a mitogen. It is a known gene affecting the heart rate [Silva et al., 2007]. The variant allele G in HTR3B (rs17116117) shows significantly in-crease risk compared with common homozygote genotype, especially among heterozygote
Table 4.1: Genes of the Genome Showing the Strongest Association
Gene Chromosome dbSNP ID Function Trend
P-value
Genotypic P-value
1 rs10494787 4.69E-02 3.65E-22
1 rs825148 3.25E-10 2.50E-101
2 rs1870340 3.30E-08 4.79E-36
3 rs804980 1.43E-03 5.64E-10
4 rs16837871 3.27E-26 1.80E-41
4 rs1553460 1.22E-13 1.29E-62
LOC100129858 4 rs6840033 Intron 1.64E-12 8.94E-23
5 rs4867173 2.28E-08 1.72E-08
5 SNP A-2171701 2.67E-02 4.46E-08
6 rs4131463 6.25E-14 4.90E-89
6 rs10499044 3.01E-15 5.47E-24
7 rs193837 2.97E-04 4.09E-27
RPL18P4 7 rs1528356 Intron 5.81E-12 2.96E-133
CHRM2* 7 rs7800093 Intron 1.59E-06 6.25E-44
KCNB2* 8 rs11782342 Intron 9.20E-04 6.59E-08
9 rs7864098 9.20E-01 5.12E-10
9 rs17797701 1.07E-03 2.48E-52
9 rs488101 4.50E-07 2.19E-09
10 rs11005510 2.36E-10 3.65E-23
OTOG 11 rs11024327 Intron 6.61E-07 4.36E-08
HTR3B* 11 rs17116117 Intron 5.07E-49 2.70E-48
12 rs10843660 1.90E-32 1.04E-69
CHST11 12 rs11112069 Intron 4.54E-03 6.70E-11
12 rs4765066 8.52E-10 2.18E-10
13 rs17667894 5.41E-21 3.70E-40
SIP1 14 rs8011855 Intron 3.35E-03 1.23E-13
RHOJ 14 rs1957779 nearGene-5 2.34E-05 5.39E-12
14 rs6574988 2.00E-07 1.03E-06
15 rs2865199 8.24E-10 3.68E-12
16 rs16955238 3.88E-06 3.61E-41
17 SNP A-1948953 6.31E-06 1.81E-13
17 rs7217721 3.80E-04 2.47E-09
*Denotes the gene or SNP has been found in published document.
The variant in rs2820037 is significantly associated with hypertension as the previous study described [The Wellcome Trust Case Control Consortium, 2007], [Ehret et al., 2008].
The SNP rs11782342 has a significant increase in risk among heterozygote variants: odds ratio [95% CI] = 1.41[1.24-1.60].
GAB1 (GRB2-associated binding protein 1) encodes the protein which is a member of the IRS1-like multisubstrate docking protein family. The protein is an important me-diator of branching tubulogenesis and plays a central role in cellular growth response,
Table 4.2: Detection of SNPs with the Strongest Association rs10494787 G 0.69[0.57-0.83] 14.09[6.45-30.78] 0.068 0.079 rs825148 C 0.05[0.02-0.10] Inf[NaN-Inf] 0.041 0.078 rs1870340 G 0.31[0.20-0.49] 114.32[15.88-822.97] 0.021 0.044 rs804980 A 0.91[0.80-1.03] 2.02[1.60-2.54] 0.217 0.246 rs16837871 A 0.36[0.31-0.42] 0.79[0.58-1.08] 0.183 0.101 rs1553460 T 0.61[0.53-0.69] 2.82[2.37-3.34] 0.291 0.369 rs6840033 T 0.52[0.45-0.59] 0.88[0.69-1.12] 0.236 0.174 rs4867173 T 1.48[1.30-1.68] 1.22[0.75-1.98] 0.132 0.171 SNP A-2171701 T 0.93[0.80-1.07] 3.18[2.09-4.84] 0.117 0.132 rs4131463 C 0.09[0.05-0.16] 116.72[28.89-471.54] 0.037 0.081 rs10499044 C 0.44[0.37-0.51] 0.97[0.67-1.42] 0.134 0.081 rs193837 C 0.74[0.62-0.88] 10.38[5.90-18.24] 0.084 0.107 rs1528356 G 0.00[0.00-0.03] 27.77[15.09-51.08] 0.057 0.104 rs7800093 G 0.02[0.00-0.11] 53.00[12.98-216.38] 0.017 0.036 rs11782342 A 0.97[0.86-1.10] 1.98[1.56-2.53] 0.226 0.255 rs7864098 A 0.75[0.64-0.88] 3.62[2.20-5.97] 0.090 0.091 rs17797701 G 0.01[0.00-0.08] 28.04[10.24-76.79] 0.024 0.038 rs488101 C 0.68[0.60-0.77] 0.74[0.62-0.88] 0.384 0.334 rs11005510 A 0.01[0.00-0.10] Inf[NaN-Inf] 0.017 0.003 rs11024327 A 1.44[1.27-1.63] 1.14[0.81-1.59] 0.172 0.212 rs17116117 G 3.76[3.13-4.52] 1.77[0.11-28.34] 0.032 0.101 rs10843660 T 0.31[0.27-0.35] 0.53[0.45-0.62] 0.430 0.303 rs11112069 A 0.88[0.77-1.00] 2.21[1.71-2.85] 0.183 0.207 rs4765066 A 1.55[1.36-1.76] 1.22[0.78-1.92] 0.129 0.173 rs17667894 G 0.02[0.01-0.07] 1.62[0.58-4.46] 0.035 0.005 rs8011855 A 0.88[0.74-1.05] 8.53[4.34-16.77] 0.069 0.086 rs1957779 A 1.69[1.46-1.96] 1.44[1.21-1.72] 0.474 0.515 rs6574988 T 1.45[1.26-1.67] 1.63[0.83-3.20] 0.090 0.122 rs2865199 C 0.21[0.12-0.35] Inf[NaN-Inf] 0.019 0.005 rs16955238 C 0.22[0.13-0.35] Inf[NaN-Inf] 0.022 0.042 SNP A-1948953 A 0.99[0.88-1.12] 0.35[0.26-0.48] 0.302 0.262 rs7217721 C 1.05[0.84-1.30] 15.88[4.85-52.01] 0.037 0.053
transformation and apoptosis. Carriers of the variant T of GAB1 (rs300916) has a signifi-cantly lower risk of hypertension compared with individuals with the common homozygote genotype: odds ratio [95% CI] for heterozygotes 0.81 [0.72-0.92] and for homozygotes 0.67 [0.56-0.80]. Nakaoka has proved that the relationship between GAB1 and hypertrophic cardiomyopathy [Nakaoka et al., 2003], and hypertension can result in hypertrophic car-diomyopathy.
BCAT1 (branched chain aminotransferase 1, cytosolic) encodes the cytosolic form of
Table 4.3: Genes of the Genome Showing Moderate Association
Gene Chromosome dbSNP ID Function Trend
P-value
Genotypic P-value
NEGR1 1 rs10889923 Intron 1.13E-01 2.03E-06
1 rs1896250 3.84E-04 5.08E-07
1 rs12729977 6.25E-01 9.05E-06
1 rs2820026 6.70E-05 3.96E-06
1 rs9428826 1.21E-04 1.95E-06
1 rs2790622 7.96E-05 8.58E-07
1 rs2820037* 8.10E-05 7.78E-07
1 rs2820038 7.25E-05 9.26E-07
1 rs2820046 8.35E-05 1.12E-06
CREG2 2 rs4850969 Intron 1.50E-01 2.00E-06
PRKCI 3 rs2140825 Intron 4.93E-02 5.01E-06
GAB1* 4 rs300916 Intron 2.49E-06 1.45E-05
LOC100128588 6 rs1935683 Intron 9.33E-05 7.29E-06
CNBD1 8 rs7825717 Intron 9.36E-01 9.28E-07
ZHX2 8 rs10095188 Intron 1.27E-02 9.48E-06
8 rs4242382 8.96E-06 3.86E-05
8 rs11166882 9.58E-06 5.03E-05
BCAT1* 12 rs7961152 Intron 2.86E-06 1.41E-05
MYBPC1* 12 rs11110912 Intron 8.12E-06 1.84E-05
15 rs921535 1.63E-05 5.47E-06
LOC100132798* 15 rs2398162 Intron 2.13E-06 1.44E-06
YWHAE 17 rs16945811 Intron 5.54E-07 2.24E-06
17 rs17201619 3.58E-06 4.69E-06
ZNF236 18 rs4890866 Intron 2.04E-02 5.34E-06
SEC23B 20 rs1022684 nearGene-5 2.36E-06 4.19E-06
*Denotes the gene or SNP has been found in published document.
sential for cell growth. Hypertension can cause atherosclerosis, furthermore, BCAT has been implicated in the pathogenesis of atherosclerosis [Coles et al., 2009]. Carriers of the variant A of BCAT1 (rs7961152) has a significantly higher risk of hypertension com-pared with individuals with the common homozygote genotype: odds ratio [95% CI] for heterozygotes 1.17 [1.03-1.34] and for homozygotes 1.49 [1.26-1.76] [The Wellcome Trust Case Control Consortium, 2007].
MYBPC1 (rs11110912). Carriers of the variant G of MYBPC1 (rs11110912) has a significantly higher risk of hypertension compared with individuals with the common homozygote genotype: odds ratio [95% CI] for heterozygotes 1.33 [1.18-1.51] and for homozygotes 1.34 [0.97-1.86] [The Wellcome Trust Case Control Consortium, 2007]. In the previous study, MYBPC1 is also related to hypertrophic cardiomyopathy [Konno et al., 2003].
LOC100132798 is similar to hCG1774772. Carriers of the variant G of LOC100132798 (rs2398162) has a significantly higher or lower risk of hypertension compared with
individ-uals with the common homozygote genotype: odds ratio [95% CI] for heterozygotes 24.33 [3.22-183.63] and for homozygotes 0.75 [0.59-0.95] [The Wellcome Trust Case Control Consortium, 2007].
SEC23B (Sec23 homolog B (S. cerevisiae)) encodes the protein which is a member of the SEC23 subfamily of the SEC23/SEC24 family. The encoded protein has similarity to yeast Sec23p component of COPII. COPII is the coat protein complex responsible for vesicle budding from the ER. The function of this gene product has been implicated in cargo selection and concentration. Subjects with the variant T of SEC23B (rs1022684) shows significantly reduced risk compared with common homozygote genotype: odds ratio [95% CI] for heterozygotes 0.70 [0.58-0.83] and for homozygotes 0.21 [0.06-0.69].
Table 4.4: Detection of SNPs with Moderate Association
dbSNP ID Minor rs10889923 C 1.18[1.04-1.34] 0.77[0.64-0.92] 0.410 0.394 rs1896250 A 1.41[1.24-1.60] 1.21[1.01-1.45] 0.379 0.414 rs12729977 C 1.22[1.08-1.39] 0.83[0.69-1.00] 0.402 0.397 rs2820026 T 1.39[1.22-1.58] 0.97[0.65-1.44] 0.138 0.167 rs9428826 T 1.40[1.23-1.59] 0.93[0.64-1.35] 0.140 0.168 rs2790622 C 1.41[1.24-1.60] 0.90[0.61-1.33] 0.141 0.170 rs2820037 T 1.41[1.24-1.60] 0.89[0.60-1.32] 0.141 0.170 rs2820038 T 1.41[1.24-1.60] 0.90[0.61-1.34] 0.141 0.170 rs2820046 A 1.40[1.23-1.60] 0.90[0.61-1.33] 0.141 0.170 rs4850969 T 1.02[0.89-1.18] 0.08[0.02-0.32] 0.113 0.104 rs2140825 C 1.12[0.99-1.27] 0.71[0.59-0.87] 0.399 0.381 rs300916 T 0.81[0.72-0.92] 0.67[0.56-0.80] 0.406 0.359 rs1935683 C 0.73[0.65-0.83] 0.95[0.69-1.31] 0.198 0.167 rs7825717 C 1.14[0.97-1.33] 0.00[0.00-NaN] 0.082 0.081 rs10095188 C 1.02[0.90-1.16] 0.45[0.31-0.63] 0.185 0.165 rs4242382 A 0.73[0.63-0.84] 0.64[0.35-1.18] 0.125 0.097 rs11166882 T 0.64[0.35-1.18] 0.68[0.54-0.85] 0.285 0.244 rs7961152 A 1.17[1.03-1.34] 1.49[1.26-1.76] 0.413 0.461 rs11110912 G 1.33[1.18-1.51] 1.34[0.97-1.86] 0.165 0.200 rs921535 C 1.38[1.21-1.57] 1.07[0.70-1.63] 0.141 0.173 rs2398162 G 24.33[3.22-183.63] 0.75[0.59-0.95] 0.260 0.218 rs16945811 A 1.48[1.27-1.72] 1.50[0.70-3.19] 0.074 0.102 rs17201619 A 0.71[0.60-0.85] 0.19[0.06-0.63] 0.079 0.055 rs4890866 G 1.07[0.95-1.20] 0.61[0.49-0.77] 0.322 0.300 rs1022684 T 0.70[0.58-0.83] 0.21[0.06-0.69] 0.078 0.054
Figure 4.3: Genome-wide Manhattan Plot for Hypertension on Multiple SNPs-Based by Chi-square Test
Multiple SNPs association
According to the interactions of SNPs within the strongest and moderate association, side effects are also siginificant if main effects are associated with disease. Consequently, we do not focus on known and obvious interactions, we are interested in SNPs that are usually ignored, namely, we focus on the interactions of SNPs without single SNP associations we found before. In addition to this, we can apply filterable method as mentioned in chapter 3, setting λAB = 1.75, fA = 0.2, fB = 0.2 by conservative rule due to the estimate ˆλAB in interactions of SNPs within the strongest and moderate association are pretty high (even ˆλAB = 6). Thus we can reduce computation time about (1
-C226108
C2406088) = 99.59% by p-value is higher than 1 × 10−1 in single association, i.e. we set ξ1 = 2.7 (α = 0.1) due to our tolerable loss of power is under 1%. Of course, adjusting the threshold ξ1 repeatedly for the methodology as mentioned in chapter 3 can find the threshold ξ1 as exact as possible. Consequently, the computation time would be improved as possible.
In the beginning, we narrowed down the target SNPs for less computation time by p-value between 1 × 10−4 and 1 × 10−5 in single association. By figure 4.3, we listed interactions within chromosome at table 4.5 with 1 × 10−110≤ p-value ≤ 1 × 10−125, and figure 4.4 shows the relation of p-value between single SNP and paired SNPs association.
The SNPs rs2091244, rs2177686, rs17073046 all locate on the gene MAGI1. MAGI1
Figure 4.4: The Relation of P-value Between Single SNP & Paired SNPs Association for Hypertension
(membrane associated guanylate kinase, WW and PDZ domain containing 1) encodes the protein which is a member of the membrane-associated guanylate kinase homologue (MAGUK) family.The product of this gene may play a role as scaffolding protein at cell-cell junctions. To date, we just know that MAGI1 is important for vascular endothelial-cadherin-dependent Rap1 activation upon cell-cell contact [Sakurai et al., 2006], however, we cannot connect it with hypertension.
GAB1 and BCAT1 not only have been found in the single SNP association we men-tioned before but also have been proved by previous study. However, some interactions on genes C10orf72, C10orf128, LOC728883 or not identify genes have not yet been proposed and proven from the biological aspect.
Table 4.5: Detection of Multiple SNPs-Based Association
Chromosome dbSNP ID 1 (Gene 1)
(MAGI1*) 1.47E-115 9.80E-05 1.77E-04 6.55
3 rs2091244
(MAGI1*)
rs17073046
(MAGI1*) 3.11E-117 9.80E-05 1.22E-04 6.58
4 rs300915
(GAB1*)
rs300913
(GAB1*) 4.00E-112 5.06E-05 4.71E-05 6.44 5 rs1490800 rs1490796 3.09E-114 9.94E-05 7.55E-05 5.95 5 rs1490800 rs1490795 9.17E-115 9.94E-05 7.75E-05 5.95 5 rs1490796 rs1490795 1.06E-114 7.55E-05 7.75E-05 5.96 10 rs12269023
(C10orf72)
rs7097933
(C10orf72) 1.54E-112 3.72E-05 3.46E-05 6.77 10 rs2725181
(C10orf128)
rs2725190
(LOC728883) 5.47E-111 7.86E-05 1.58E-04 8.67 12 rs11613673
(BCAT1*)
rs12424348
(BCAT1*) 4.83E-120 6.95E-05 1.49E-04 10.28 12 rs7300456 rs1452237 3.97E-113 1.65E-05 1.91E-05 6.92 12 rs4761100 rs4761102 5.44E-116 2.97E-05 2.33E-05 7.66 20 rs2424430 rs431904 2.53E-111 1.18E-05 3.65E-05 8.15
*Denotes the gene or SNP has been found in published document.
Chapter 5 Conclusion
According to the results in table 3.5, table 3.6, and the real data, the loss of power is reasonable and tolerable when λAB is large enough or the allele frequency is not too small. Each pair of SNPs association has an unknown λAB originally, but estimate all λAB is unusable because our major work is to find out a reasonable threshold by only one λAB and other parameters. We found that ˆλAB within the strongest or moderate associations are quite large, such as 6.0 or 7.6, but we cannot promise that λAB for all existing associations are large, too. That is the reason why we use more conservative and robust rule as λAB = 1.75 in this study. We can reduce computation time about
• 99.04% = (1 − C239762
C2406088), loss of power = 0.2612%, when ξ1 = 2.07 (α = 0.15)
• 99.59% = (1 − C226108
C2406088), loss of power = 0.8224%, when ξ1 = 2.7 (α = 0.1)
• 99.77% = (1 − C219424
C2406088), loss of power = 1.6915%, when ξ1 = 3.17 (α = 0.075) Analyzing the data with this approach, which imitates WTCCC of hypertension, we have detected parts of known genes or SNPs, such as CHRM2 (rs7800093), KCNB2 (rs11782342), HTR3B (rs17116117), rs2820037, GAB1 (rs300916, rs300915, rs300913), BCAT1 (rs7961152, rs11613673, rs12424348), MYBPC1 (rs11110912), LOC100132798 (rs2398162), MAGI1 (rs2091244, rs2177686, rs17073046). Nevertheless, those other un-knowns, such as rs825148, rs1553460, LOC100129858 (rs6840033), rs4131463, RPL18P4 (rs1528356), rs17797701, OTOG (rs11024327), rs10843660, CHST11 (rs11112069), SIP1 (rs8011855), RHOJ (rs1957779) are worthy of digging for statistical replication and bio-logical explanation in the future. Furthermore, the associations of higher order are also our ultimate goal for finding the susceptibility for complex human diseases, for instance, hypertension and type 2 diabetes.
some SNPs’ associations are clearly quite different in these two figures. Thus the extension for no model assumption may be more accurate and informative (single association test uses genotypic test instead of trend test).
We have not considered the dominant or recessive model in the method and analysis.
In general, the models for most of SNPs are still unknown, integrate information (consider the dominant or recessive model additionally) from every models and revise our method is a part of future work. Using this method to calculate the loss of power and use ECM algorithm to find suitable parameters may provide a good guidance to threshold selection.
Bibliography
N. Risch and K. Merikangas. The future of genetic studies of complex human diseases.
Science, 273:1516–1517, 1996.
C. E. Bonferroni. Teoria statistica delle classi e calcolo delle probabilit‘a. Pubblicazioni del R Istituto Superiore di Scienze Economiche e Commerciali di Firenze, 8:3–62, 1936.
S. Holm. A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6:65–70, 1979.
Y. Hochberg. A sharper Bonferroni procedure for multiple tests of significance.
Biometrika, 75:800–802, 1988.
Y. Benjamini and Y. Hochberg. Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal Statistical Society B, 57:
289–300, 1995.
P. Armitage. Tests for linear trends in proportions and frequencies. Biometrics, 11:
375–386, 1955.
R. A. Fisher. On the Interpretation of q2 from Contingency Tables, and the Calculation of P. Journal of the Royal Statistical Society, 85:87–94, 1922.
K. Pearson. On the criterion that a given system of deviations from the probable in the case of a correlated system of variables is such that it can be reasonably supposed to hove arisen from random sampling. Phios. Mag. Ser., 50:157–175, 1900.
X. L. Meng and D. B. Rubin. Maximum likelihood estimation via the ECM algorithm:
A general framework. Biometrika, 80:267–278, 1993.
The Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature, 447:661–678, 2007.
A. J. Hautala, M. P. Tulppo, A. M. Kiviniemi, T. Rankinen, C. Bouchard, T. H.
M’akikallio, and H. V. Huikuri. Acetylcholine receptor m2 gene variants, heart rate
L. Zhang, A. Hu, H. Yuan, L. Cui, G. Miao, X. Yang, L. Wang, J. Liu, X. Liu, S. Wang, Z. Zhang, L. Liu, R. Zhao, and Y. Shen. A missense mutation in the chrm2 gene is associated with familial dilated cardiomyopathy. Circulation Research, 102:1426–1432, 2008.
R. S. Vasan, M. G. Larson, J. Aragam, T. J. Wang, G. F. Mitchell, S. Kathiresan, C. Newton-Cheh, J. A. Vita, M. J. Keyes, C. J. O’Donnell, D. Levy, and E. J. Bjamin. Genome-wide association of echocardiographic dimensions, brachial artery en-dothelial function and treadmill exercise responses in the Framingham Heart Study.
BMC Medical Genetics, 8 Suppl 1:S2, 2007.
Gustavo J. J. Silva, Alexandre C. Pereira, Eduardo M. Krieger, and Jos’e E. Krieger. Ge-netic mapping of a new heart rate QTL on chromosome 8 of spontaneously hypertensive rats. BMC Medical Genetics, 8:17, 2007.
G. B. Ehret, A. C. Morrison, A. A. O’Connor, M. L. Grove, L. Baird, K. Schwander, A. Weder, R. S. Cooper, D. C. Rao, S. C. Hunt, E. Boerwinkle, and A. Chakravarti.
Replication of the Wellcome Trust genome-wide association study of essential hyper-tension: the Family Blood Pressure Program. European Journal of Human Genetics, 16:1507–1511, 2008.
Y. Nakaoka, K. Nishida, Y. Fujio, M. Izumi, K. Terai, Y. Oshima, S. Sugiyama, S. Mat-suda, S. Koyasu, K. Yamauchi-Takihara, T. Hirano, I. Kawase, and H. Hirota. Activa-tion of gp130 transduces hypertrophic signal through interacActiva-tion of scaffolding/docking protein Gab1 with tyrosine phosphatase SHP2 in cardiomyocytes. Circulation Research, 93:221–229, 2003.
S. J. Coles, P. Easton, H. Sharrod, S. M. Hutson, J. Hancock, V. B. Patel, and M. E.
Conway. S-Nitrosoglutathione inactivation of the mitochondrial and cytosolic BCAT proteins: S-nitrosation and S-thiolation. Biochemistry, 48:645–656, 2009.
T. Konno, M. Shimizu, H. Ino, T. Matsuyama, M. Yamaguchi, H. Terai, K. Hayashi, M. Mabuchi, T. amd Kiyama, K. Sakata, T. Hayashi, M. Inoue, T. Kaneda, and H. Mabuchi. A novel missense mutation in the myosin binding protein-C gene is respon-sible for hypertrophic cardiomyopathy with left ventricular dysfunction and dilation in elderly patients. Journal of the American College of Cardiology, 41:781–786, 2003.
A. Sakurai, S. Fukuhara, A. Yamagishi, K. Sako, Y. Kamioka, M. Masuda, Y. Nakaoka, and N. Mochizuki. MAGI-1 is required for Rap1 activation upon cell-cell contact and for enhancement of vascular endothelial cadherin-mediated cell adhesion. Molecular Biology of the Cell, 17:966–976, 2006.
T. Becker and M. Knapp. Maximum-likelihood estimation of haplotype frequencies in nuclear families. Genet Epidemiol, 27:21–32, 2004.
D. Sasieni. From genotypes to genes: doubling the sample size. Biometrics, 53:1253–1261, 1997.
Mark C. K. Yang. Introduction to Statistical Methods in Modern Genetics. Gordon and Breach Science Publishers, 2000.
K. Hao, X. Xu, N. Laird, X. Wang, and X. Xu. Power Estimation of Multiple SNP Association Test of Case-Control Study and Application. Genetic Epidemiology, 26:
22–30, 2004.
National Center for Biotechnology Information, National Library of Medicine, and Na-tional Institutes of Health. Entrez gene. http://www.ncbi.nlm.nih.gov/sites/
entrez?db=gene, June 2009.