Combining an Evolution-guided Clustering Algorithm and Haplotype-based LRT in Family Association Studies
© Lee et al; licensee BioMed Central Ltd. 2011
Received: 3 October 2010
Accepted: 19 May 2011
Published: 19 May 2011
With the completion of the international HapMap project, many studies have been conducted to investigate the association between complex diseases and haplotype variants. Such haplotype-based association studies, however, often face two difficulties; one is the large number of haplotype configurations in the chromosome region under study, and the other is the ambiguity in haplotype phase when only genotype data are observed. The latter complexity may be handled based on an EM algorithm with family data incorporated, whereas the former can be more problematic, especially when haplotypes of rare frequencies are involved. Here based on family data we propose to cluster long haplotypes of linked SNPs in a biological sense, so that the number of haplotypes can be reduced and the power of statistical tests of association can be increased.
In this paper we employ family genotype data and combine a clustering scheme with a likelihood ratio statistic to test the association between quantitative phenotypes and haplotype variants. Haplotypes are first grouped based on their evolutionary closeness to establish a set containing core haplotypes. Then, we construct for each family the transmission and non-transmission phase in terms of these core haplotypes, taking into account simultaneously the phase ambiguity as weights. The likelihood ratio test (LRT) is next conducted with these weighted and clustered haplotypes to test for association with disease. This combination of evolution-guided haplotype clustering and weighted assignment in LRT is able, via its core-coding system, to incorporate into analysis both haplotype phase ambiguity and transmission uncertainty. Simulation studies show that this proposed procedure is more informative and powerful than three family-based association tests, FAMHAP, FBAT, and an LRT with a group consisting exclusively of rare haplotypes.
The proposed procedure takes into account the uncertainty in phase determination and in transmission, utilizes the evolutionary information contained in haplotypes, reduces the dimension in haplotype space and the degrees of freedom in tests, and performs better in association studies. This evolution-guided clustering procedure is particularly useful for long haplotypes containing linked SNPs, and is applicable to other haplotype-based association tests. This procedure is now implemented in R and is free for download.
High-density sets of SNPs, especially haplotypes, have been used widely in genetic research to explore possible association with complex diseases. Haplotypes are considered to be the biological units containing more information about transmission, and thus may be better biomarkers to use in examining the disease susceptible region. However, haplotype phase is often unknown when only genotype data are observed. This linkage phase ambiguity often leads to large degrees of freedom in statistical tests, and may result in estimation of many haplotypes with rare frequencies. Collection of family genotype data may help in determination of haplotype phase if information from other family members can be incorporated and cross-referenced. Additionally, the use of family data can avoid spurious association arising from population admixture. Nevertheless, the statistical analysis of family data may not be straightforward. For instance, the nonparametric transmission disequilibrium test (TDT) and other similar tests utilize the transmitted and non-transmitted alleles (or haplotypes) to detect association. For uncertainty both in transmission and in phase, most procedures adopt an expectation-maximization (EM) algorithm in computation of score statistics or likelihood functions, such as FBAT  and FAMHAP [2, 3]. These methods infer haplotypes based on nuclear families, where any number of children and configuration of missing genotype data are allowed. FBAT considers score statistics for the number of risk haplotypes among affected offspring; whereas FAMHAP adopts a likelihood approach to estimate haplotype frequencies and to test for association based on nuclear family members or unrelated individuals.
As for addressing the problems that result when a large number of different haplotypes are involved in analysis and when certain haplotypes occur with very low frequency, several approaches have been adopted. Some studies have deleted haplotypes with small estimated frequencies [4, 5], and some have combined these rare haplotypes to form a new group. Although these procedures can reduce the total number of parameters and the inflated variation due to small frequencies, they risk information loss due to the arbitrary deletion or grouping of rare haplotypes. In contrast, several procedures have adopted an approach which uses a measure of evolutionary relationship to cluster rare haplotypes in case-control studies [6–10]. These clustering algorithms define a core set of haplotypes that are considered "ancient," where being ancient is approximated by being more frequent [7–10]. Once the core set is determined, rare haplotypes are then clustered with their ancestors as well as denoted by them. For family data, we want to employ such a clustering scheme, so that the statistical tests can be conducted more efficiently.
Under the assumption of random mating and with the use of transmitted and non-transmitted haplotypes from parents, we construct for family data a core set of haplotypes based on estimates of haplotype frequencies. In the following sections, we start with the notation used in Becker and Knapp , conduct the clustering procedure for family data based on frequency estimates from FAMHAP, determine the pair of transmitted and non-transmitted core-coding haplotypes, and introduce the assignment of coding to be used later in our tests of disease association. The uncertainty in phase explanation, in transmission status, and in core representation associated with each genotype will be expressed with weights. Unlike the analysis for case-control studies, these weights are essential for further statistical analysis of pedigree data. We next adopt this transformed data in a likelihood ratio test (LRT) of association, and call it LRT-C. This LRT-C modifies the original test in FAMHAP in terms of the core haplotypes. To evaluate if the proposed clustering approach can identify correctly the true core haplotypes and to demonstrate the advantage of LRT-C accrued by inclusion of an evolutionary interpretation, we conduct simulation studies by applying a coalescent-based whole genome simulator GENOME . For the purpose of comparison with other tests, we consider the original LRT in FAMHAP, a score test in FBAT, and a naïve LRT with all rare haplotypes clustered to form a new group (LRT-G).
It is worth noting that, under the null hypothesis of no association, is the intersection of all for l' = 1,...,n i in the i-th family. Consequently, the estimation of haplotype frequency is independent of who the proband is. In contrast, when the gene is associated with the disease, the haplotype explanation set and the estimations of the transmitted and non-transmitted haplotype frequencies will depend on the ascertained proband. Therefore, throughout this paper, we rearrange the order of children such that the first child is always the first affected one in each family. In the following, we use FAMHAP to estimate the configurations of haplotypes and their corresponding frequencies based on likelihoods. In fact, at this stage of computation, other public software programs like FBAT and Transmit  can provide stable estimates as well.
Step 1 Clustering haplotypes
With the frequency approximating the "age" of haplotypes, a cladistic clustering approach is conducted based on their evolutionary relation. Similar to the haplotype clustering method for case-control studies in Tzeng , here we identify first the ancestor (core) and descendent haplotypes based on family data, and then cluster the descendents with their ancestors for dimension reduction.
where (∏(m)) t is the transpose of ∏(m). Detailed explanation is provided in Additional file 1.
Note that the above derivations do not require the information of haplotype phase for each family member. In other words, any software which provides estimates of haplotype frequencies can be applied at this stage. However, we prefer FAMHAP because it also computes for each family the compatible transmitted and non-transmitted haplotypes, along with the weights, which will be utilized in the following steps of recoding and testing.
Step 2 Recoding transmission and non-transmission haplotypes for analysis
Step 3 Likelihood ratio test with clustered haplotypes (LRT-C)
The γ j.Tr and γ u.Tr stand for frequencies in the new representation (2) for the transmitted haplotypes (j,u), and γ k.NTr and γ v.NTr are the modified frequencies for the non-transmitted haplotypes (k,v). The function s is the number of transmission patterns compatible with the genotype of the l'-th child c l' , as defined in Becker and Knapp's paper. Note that the value of will not differ no matter what haplotypes, core or not, are being investigated. That is, there is no need to recompute s with respect to the core system, and its value derived when specifying the haplotype explanation set in the beginning of this procedure can be used directly here.
To evaluate the performance of the likelihood ratio test with clustered haplotypes in family studies, we conduct simulations to first examine the reconstruction and identification of core haplotypes, and next to evaluate the impact of this clustering scheme on the likelihood ratio test. The results are compared with three family-based association methods, FAMHAP , FBAT , and finally an LRT using a naïve group composed exclusively of rare haplotypes (LRT-G).
Sampling scheme for simulations
Penetrance values for simulation settings
r = 3
p = 0.1
p = 0.25
p = 0.5
r = 2.5
p = 0.1
p = 0.25
p = 0.5
r = 2
p = 0.1
p = 0.25
p = 0.5
Identification of core haplotypes and tests of association
Number are the power of four family-based association tests at 5% significance level with N = 200
r = 3
p = 0.1
p = 0.25
p = 0.5
r = 2.5
p = 0.1
p = 0.25
p = 0.5
r = 2
p = 0.1
p = 0.25
p = 0.5
Type I errors of the four family-based association tests at the 5% significance level
N = 200
Performance evaluation under population admixture
Numbers are the power of four family-based association tests for population stratification data at the 5% significance level with N = 200
r = 3
r = 2.5
r = 2
In this paper, we have constructed a family-based association test using clustered haplotypes. The four key steps are: (1) to determine the core set on the basis of haplotype frequencies, (2) to perform the clustering procedure based on a haplotype cladogram, (3) to represent the rare haplotype frequencies in terms of the revised core frequencies, and (4) to incorporate the phase ambiguity, transmission uncertainty, and core-representation variability via likelihood weights. Our simulations show that both haplotype reconstruction and core identification perform well with more than 91% accuracy for cases where number of families N ≥ 100. In addition, the results show great improvement in test power, as compared to the original FAMHAP, FBAT, and LRT-G. Our proposed procedure is useful for long haplotypes containing many SNPs in LD. If the SNPs are not in LD, then it is not appropriate to consider haplotypes as the unit for analysis. In addition, if the haplotypes are of short length, then both the dimensionality and phase ambiguity will not be hard to handle. We have also embedded this clustering algorithm in the likelihood ratio test under FAMHAP, and this program can be downloaded freely from the author's website http://homepage.ntu.edu.tw/~ckhsiao/download(en).html.
One issue that merits discussion concerns the number of haplotypes in the core set. As a rule of thumb, we selected the several leading haplotypes with 0.9 cumulative frequency. This choice is somewhat arbitrary. In fact, there is a trade-off between the increase in information (represented via frequency) and the reduction in dimensionality. A possible alternative, depending on the sample size and number of dimensions under consideration, would be to use Shannon's information with a penalty function. This criterion works by finding l haplotypes such that Shannon's net information reaches its maximum. This criterion, however, is sensitive with respect to sample size. When sample size gets large, the penalty decreases faster than the entropy term, and thus results in inclusion of all haplotypes in the core set, even those with small frequencies. In other words, this criterion does not effectively reduce dimensionality when large sample size prevails.
There are several potential applications for the association test presented here. In many association tests, the chi-square approximation can be poor due to the existence of many haplotypes and/or rare haplotypes. Our clustering procedure may improve the performance of such statistical methods. Although we only demonstrate its impact on the likelihood ratio test, we believe other tests would benefit from this clustering procedure as well. For instance, after the haplotype phase and transmission status are identified and recoded via the core for each family, the tests in TRANSMIT or FBAT or other kinds of haplotype inference [19–22] can be modified. Another application is to use the clustered haplotypes in regression analysis to incorporate the environmental influence on quantitative traits [23, 24]. Because FAMHAP provides for each family and its individual family members the list of all possible haplotype explanations with corresponding likelihoods, this information can be utilized in further analyses. In addition, it would be interesting to extend this clustering approach to data containing both independent and related individuals. A mixture of family-based and population-based data are sometimes considered in meta-genetic analysis to enhance the power of an association study. Applying this clustering technique can further reduce the dimension of parameters and achieve a larger power in detection of genetic association with common diseases. Finally, we would like to point out that we consider in this paper only simulations from one single population or from a mixture of two populations. Though the results look promising, other scenarios are warrant for further investigation.
For family genotype data, we consider an evolution-guided clustering tool that clusters rare haplotypes in order to achieve dimensional reduction, and a parametric likelihood ratio test that accounts for the uncertainty associated with transmission phase. This procedure is able to preserve biological information and to improve statistical testing power. Simulation studies of long haplotypes with SNPs in LD show that the proposed likelihood ratio test with clustered haplotypes (LRT-C) outperforms FAMHAP, FBAT, and a naïve LRT-G.
MHL was supported in part by NSC 99-2118-M-113-001. JYT was supported by NIH grants R01MH84022 and P01CA142538. SYH, and CKH were supported in part by NSC 97-2314-B-002-040-MY3.
- Horvath S, Xu X, Lake SL, Silverman EK, Weiss ST, Laird NM: Family-based tests for association haplotypes with general phenotype data: application to asthma genetics. Genet Epidemiol. 2004, 26: 61-69. 10.1002/gepi.10295.View ArticlePubMedGoogle Scholar
- Becker T, Knapp M: Maximum-likelihood estimation of haplotype frequencies in nuclear families. Genet Epidemiol. 2004, 27: 21-32. 10.1002/gepi.10323.View ArticlePubMedGoogle Scholar
- Herold C, Becker T: Genetic association analysis with FAMHAP: a major program update. Bioinformatics. 2009, 25: 134-136. 10.1093/bioinformatics/btn581.View ArticlePubMedGoogle Scholar
- Schaid DJ, Rowland CM, Tines DE, Jacobson RM, Poland GA: Score tests for association between traits and haplotypes when linkage phase is ambiguous. Am J Hum Genet. 2002, 70: 425-434. 10.1086/338688.PubMed CentralView ArticlePubMedGoogle Scholar
- Huang BE, Amos CI, Lin DY: Detecting haplotype effects in genomewide association studies. Genet Epidemiol. 2007, 31 (8): 803-812. 10.1002/gepi.20242.View ArticlePubMedGoogle Scholar
- Tzeng JY: Evolutionary-based grouping of haplotypes in association analysis. Genet Epidemiol. 2005, 28: 220-231. 10.1002/gepi.20063.View ArticlePubMedGoogle Scholar
- Crandall KA, Templeton AR: Empirical tests of some predictions from coalescent theory with applications to intraspecific phylogeny reconstruction. Genetics. 1993, 134: 959-969.PubMed CentralPubMedGoogle Scholar
- Slatkin M, Rannala B: Estimating the age of alleles by use of intraallelic variability. Am J Hum Genet. 1997, 60: 447-458.PubMed CentralPubMedGoogle Scholar
- Seltman H, Roeder K, Devlin B: Transmission/disequilibrium test meets measured haplotype analysis: family-based association analysis guided by evolution of haplotypes. Am J Hum Genet. 2001, 68: 1250-1263. 10.1086/320110.PubMed CentralView ArticlePubMedGoogle Scholar
- Seltman H, Roeder K, Devlin B: Evolutionary-based association analysis using haplotype data. Genet Epidemiol. 2003, 25: 48-58. 10.1002/gepi.10246.View ArticlePubMedGoogle Scholar
- Liang L, Zöllner S, Abecasis GR: Genome: a rapid coalescent-based whole genome simulator. Bioinformatics. 2007, 23: 1565-1567. 10.1093/bioinformatics/btm138.View ArticlePubMedGoogle Scholar
- Clayton D: A generalization of the transmission/disequilibrium test for uncertain-haplotype transmission. Am J Hum Genet. 1999, 65: 1170-1177. 10.1086/302577.PubMed CentralView ArticlePubMedGoogle Scholar
- Tzeng JY, Wang CH, Kao JT, Hsiao CK: Regression-based association analysis with clustered haplotypes using genotypes. Am J Hum Gene. 2006, 78: 231-242. 10.1086/500025.View ArticleGoogle Scholar
- Knapp M, Becker T: Familiy-based association analysis with tightly linked markers. Hum Hered. 2003, 56: 2-9. 10.1159/000073727.View ArticlePubMedGoogle Scholar
- Laird NM, Horvath S, Xu X: Implementing a unified approach to family-based tests of association. Genet Epidemiol. 2000, 19 (Suppl 1): S36-S42.View ArticlePubMedGoogle Scholar
- Barrett JC, Fry B, Maller J, Daly MJ: Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005, 21: 263-265. 10.1093/bioinformatics/bth457.View ArticlePubMedGoogle Scholar
- Zhao H, Zhang S, Merikangas KR, Trixler M, Wildenauer DB, Sun F, Kidd KK: Transmission/disequilibrium tests using multiple tightly linked markers. Am J Hum Genet. 2000, 67: 936-946. 10.1086/303073.PubMed CentralView ArticlePubMedGoogle Scholar
- Zheng G, Freidlin B, Gastwirth JL: Robust TDT-type candidate-gene association tests. Ann Hum Genet. 2002, 66: 145-155. 10.1046/j.1469-1809.2002.00104.x.View ArticlePubMedGoogle Scholar
- Li C, Boehnke M: Haplotype association analysis for late onset diseases using nuclear family data. Genet Epidemiol. 2006, 30: 220-230. 10.1002/gepi.20139.View ArticlePubMedGoogle Scholar
- Ding XD, Simianer H, Zhang Q: A new method for haplotype inference including full-sib information. Genetics. 2007, 177: 1929-1940. 10.1534/genetics.107.079525.PubMed CentralView ArticlePubMedGoogle Scholar
- Dudbridge F: Likelihood-based association analysis for nuclear families and unrelated subjects with missing genotype data. Hum Hered. 2008, 66: 87-98. 10.1159/000119108.PubMed CentralView ArticlePubMedGoogle Scholar
- Chen Y, Li X, Li J: A novel approach for haplotype-based association analysis using family data. BMC Bioinformatics. 2010, 11 (Suppl 1): S45-10.1186/1471-2105-11-S1-S45.PubMed CentralView ArticlePubMedGoogle Scholar
- Li J, Zhou Y, Elston RC: Haplotype-based quantitative trait mapping using a clustering algorithm. BMC Bioinformatics. 2006, 7: 258-10.1186/1471-2105-7-258.PubMed CentralView ArticlePubMedGoogle Scholar
- Cai Z, Sabaa H, Wang Y, Goebel R, Wang Z, Xu J, Stothard P, Lin G: Most parsimonious haplotype allele sharing determination. BMC Bioinformatics. 2009, 10: 115-10.1186/1471-2105-10-115.PubMed CentralView ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.