A fast and flexible statistical model for large-scale population genotype data: applications to inferring missing genotypes and haplotypic phase

We present a statistical model for patterns of genetic variation in samples of unrelated individuals from natural populations. This model is based on the idea that, over short regions, haplotypes in a population tend to cluster into groups of similar haplotypes. To capture the fact that, because of recombination, this clustering tends to be local in nature, our model allows cluster memberships to change continuously along the chromosome according to a hidden Markov model. This approach is flexible, allowing for both "block-like" patterns of linkage disequilibrium (LD) and gradual decline in LD with distance. The resulting model is also fast and, as a result, is practicable for large data sets (e.g., thousands of individuals typed at hundreds of thousands of markers). We illustrate the utility of the model by applying it to dense single-nucleotide-polymorphism genotype data for the tasks of imputing missing genotypes and estimating haplotypic phase. For imputing missing genotypes, methods based on this model are as accurate or more accurate than existing methods. For haplotype estimation, the point estimates are slightly less accurate than those from the best existing methods (e.g., for unrelated Centre d'Etude du Polymorphisme Humain individuals from the HapMap project, switch error was 0.055 for our method vs. 0.051 for PHASE) but require a small fraction of the computational cost. In addition, we demonstrate that the model accurately reflects uncertainty in its estimates, in that probabilities computed using the model are approximately well calibrated. The methods described in this article are implemented in a software package, fastPHASE, which is available from the Stephens Lab Web site.     

Direct maximum parsimony phylogeny reconstruction from genotype data

Background  

Maximum parsimony phylogenetic tree reconstruction from genetic variation data is a fundamental problem in computational genetics with many practical applications in population genetics, whole genome analysis, and the search for genetic predictors of disease. Efficient methods are available for reconstruction of maximum parsimony trees from haplotype data, but such data are difficult to determine directly for autosomal DNA. Data more commonly is available in the form of genotypes, which consist of conflated combinations of pairs of haplotypes from homologous chromosomes. Currently, there are no general algorithms for the direct reconstruction of maximum parsimony phylogenies from genotype data. Hence phylogenetic applications for autosomal data must therefore rely on other methods for first computationally inferring haplotypes from genotypes.     

Evaluation of LD decay and various LD-decay estimators in simulated and SNP-array data of tetraploid potato

Key message

The number of SNPs required for QTL discovery is justified by the distance at which linkage disequilibrium has decayed. Simulations and real potato SNP data showed how to estimate and interpret LD decay.

Abstract

The magnitude of linkage disequilibrium (LD) and its decay with genetic distance determine the resolution of association mapping, and are useful for assessing the desired numbers of SNPs on arrays. To study LD and LD decay in tetraploid potato, we simulated autotetraploid genotypes and used it to explore the dependence on: (1) the number of haplotypes in the population (the amount of genetic variation) and (2) the percentage of haplotype specific SNPs (hs-SNPs). Several estimators for short-range LD were explored, such as the average r ², median r ², and other percentiles of r ² (80, 90, and 95 %). For LD decay, we looked at LD_½,90, the distance at which the short-range LD is halved when using the 90 % percentile of r ² at short range, as estimator for LD. Simulations showed that the performance of various estimators for LD decay strongly depended on the number of haplotypes, although the real value of LD decay was not influenced very much by this number. The estimator LD_½,90 was chosen to evaluate LD decay in 537 tetraploid varieties. LD_½,90 values were 1.5 Mb for varieties released before 1945 and 0.6 Mb in varieties released after 2005. LD_½,90 values within three different subpopulations ranged from 0.7 to 0.9 Mb. LD_½,90 was 2.5 Mb for introgressed regions, indicating large haplotype blocks. In pericentromeric heterochromatin, LD decay was negligible. This study demonstrates that several related factors influencing LD decay could be disentangled, that no universal approach can be suggested, and that the estimation of LD decay has to be performed with great care and knowledge of the sampled material.

     

Effectiveness of computational methods in haplotype prediction

Haplotype analysis has been used for narrowing down the location of disease-susceptibility genes and for investigating many population processes. Computational algorithms have been developed to estimate haplotype frequencies and to predict haplotype phases from genotype data for unrelated individuals. However, the accuracy of such computational methods needs to be evaluated before their applications can be advocated. We have experimentally determined the haplotypes at two loci, the N-acetyltransferase 2 gene ( NAT2, 850 bp, n=81) and a 140-kb region on chromosome X ( n=77), each consisting of five single nucleotide polymorphisms (SNPs). We empirically evaluated and compared the accuracy of the subtraction method, the expectation-maximization (EM) method, and the PHASE method in haplotype frequency estimation and in haplotype phase prediction. Where there was near complete linkage disequilibrium (LD) between SNPs (the NAT2 gene), all three methods provided effective and accurate estimates for haplotype frequencies and individual haplotype phases. For a genomic region in which marked LD was not maintained (the chromosome X locus), the computational methods were adequate in estimating overall haplotype frequencies. However, none of the methods was accurate in predicting individual haplotype phases. The EM and the PHASE methods provided better estimates for overall haplotype frequencies than the subtraction method for both genomic regions.     

Assessment of linkage disequilibrium by the decay of haplotype sharing, with application to fine-scale genetic mapping.

Linkage disequilibrium (LD) is of great interest for gene mapping and the study of population history. We propose a multilocus model for LD, based on the decay of haplotype sharing (DHS). The DHS model is most appropriate when the LD in which one is interested is due to the introduction of a variant on an ancestral haplotype, with recombinations in succeeding generations resulting in preservation of only a small region of the ancestral haplotype around the variant. This is generally the scenario of interest for gene mapping by LD. The DHS parameter is a measure of LD that can be interpreted as the expected genetic distance to which the ancestral haplotype is preserved, or, equivalently, 1/(time in generations to the ancestral haplotype). The method allows for multiple origins of alleles and for mutations, and it takes into account missing observations and ambiguities in haplotype determination, via a hidden Markov model. Whereas most commonly used measures of LD apply to pairs of loci, the DHS measure is designed for application to the densely mapped haplotype data that are increasingly available. The DHS method explicitly models the dependence among multiple tightly linked loci on a chromosome. When the assumptions about population structure are sufficiently tractable, the estimate of LD is obtained by maximum likelihood. For more-complicated models of population history, we find means and covariances based on the model and solve a quasi-score estimating equation. Simulations show that this approach works extremely well both for estimation of LD and for fine mapping. We apply the DHS method to published data sets for cystic fibrosis and progressive myoclonus epilepsy.     

Using haplotypes for the prediction of allelic identity to fine-map QTL: characterization and properties

Background

Numerous methods have been developed over the last decade to predict allelic identity at unobserved loci between pairs of chromosome segments along the genome. These loci are often unobserved positions tested for the presence of quantitative trait loci (QTL). The main objective of this study was to understand from a theoretical standpoint the relation between linkage disequilibrium (LD) and allelic identity prediction when using haplotypes for fine mapping of QTL. In addition, six allelic identity predictors (AIP) were also compared in this study to determine which one performed best in theory and application.

Results

A criterion based on a simple measure of matrix distance was used to study the relation between LD and allelic identity prediction when using haplotypes. The consistency of this criterion with the accuracy of QTL localization, another criterion commonly used to compare AIP, was evaluated on a set of real chromosomes. For this set of chromosomes, the criterion was consistent with the mapping accuracy of a simulated QTL with either low or high effect. As measured by the matrix distance, the best AIP for QTL mapping were those that best captured LD between a tested position and a QTL. Moreover the matrix distance between a tested position and a QTL was shown to decrease for some AIP when LD increased. However, the matrix distance for AIP with continuous predictions in the [0,1] interval was algebraically proven to decrease less rapidly up to a lower bound with increasing LD in the simplest situations, than the discrete predictor based on identity by state between haplotypes (IBS _hap), for which there was no lower bound. The expected LD between haplotypes at a tested position and alleles at a QTL is a quantity that increases naturally when the tested position gets closer to the QTL. This behavior was demonstrated with pig and unrelated human chromosomes.

Conclusions

When the density of markers is high, and therefore LD between adjacent loci can be assumed to be high, the discrete predictor IBS _hap is recommended since it predicts allele identity correctly when taking LD into account.     

Genome-wide distribution of linkage disequilibrium in the population of Palau and its implications for gene flow in Remote Oceania

Linkage disequilibrium (LD) between alleles on the same human chromosome results from various evolutionary processes and is thus telling about the history of populations. Recently, LD has garnered substantial interest for its value to map and fine-map disease genes. We examine the distribution of LD between short tandem repeat alleles on autosomes and sex chromosomes in the Remote Oceanic population of Palau to evaluate whether the data are consistent with a recent hypothesis about the origins of genetic variation in Palau, specifically that the population experienced extensive male-biased gene flow following initial settlement. Consistent with evolutionary theory based on effective population size, LD between X-linked alleles is stochastically greater than LD between autosomal alleles, however, small but detectable LD occurs for autosomal markers separated by substantial distances. By contrast, while Y-linked alleles experience only one-third the effective population size of X-linked alleles, their mean value for pairwise LD is only slightly larger than X-linked alleles. For a small population known to experience at least two extreme bottlenecks, 56 six-locus Y haplotypes exhibit remarkable diversity (0.96), comparable to Y diversity of Europeans, however, autosomal and X-linked markers display significantly less diversity, as measured by heterozygosity (4.1% less). Palauan Y haplotypes also fall into distinct clusters, again unlike that of Europe. We argue these data are consistent with waves of male-biased gene flow.     

Effects of introgression and recombination on haplotype structure and linkage disequilibrium surrounding a locus encoding Bymovirus resistance in barley

We present a detailed analysis of linkage disequilibrium (LD) in the physical and genetic context of the barley gene Hv-eIF4E, which confers resistance to the barley yellow mosaic virus (BYMV) complex. Eighty-three SNPs distributed over 132 kb of Hv-eIF4E and six additional fragments genetically mapped to its flanking region were used to derive haplotypes from 131 accessions. Three haplogroups were recognized, discriminating between the alleles rym4 and rym5, which each encode for a spectrum of resistance to BYMV. With increasing map distance, haplotypes of susceptible genotypes displayed diverse patterns driven mainly by recombination, whereas haplotype diversity within the subgroups of resistant genotypes was limited. We conclude that the breakdown of LD within 1 cM of the resistance gene was generated mainly by susceptible genotypes. Despite the LD decay, a significant association between haplotype and resistance to BYMV was detected up to a distance of 5.5 cM from the resistance gene. The LD pattern and the haplotype structure of the target chromosomal region are the result of interplay between low recombination and recent breeding history.     

Linkage disequilibrium mapping via cladistic analysis of phase-unknown genotypes and inferred haplotypes in the Genetic Analysis Workshop 14 simulated data

We recently described a method for linkage disequilibrium (LD) mapping, using cladistic analysis of phased single-nucleotide polymorphism (SNP) haplotypes in a logistic regression framework. However, haplotypes are often not available and cannot be deduced with certainty from the unphased genotypes. One possible two-stage approach is to infer the phase of multilocus genotype data and analyze the resulting haplotypes as if known. Here, haplotypes are inferred using the expectation-maximization (EM) algorithm and the best-guess phase assignment for each individual analyzed. However, inferring haplotypes from phase-unknown data is prone to error and this should be taken into account in the subsequent analysis. An alternative approach is to analyze the phase-unknown multilocus genotypes themselves. Here we present a generalization of the method for phase-known haplotype data to the case of unphased SNP genotypes. Our approach is designed for high-density SNP data, so we opted to analyze the simulated dataset. The marker spacing in the initial screen was too large for our method to be effective, so we used the answers provided to request further data in regions around the disease loci and in null regions. Power to detect the disease loci, accuracy in localizing the true site of the locus, and false-positive error rates are reported for the inferred-haplotype and unphased genotype methods. For this data, analyzing inferred haplotypes outperforms analysis of genotypes. As expected, our results suggest that when there is little or no LD between a disease locus and the flanking region, there will be no chance of detecting it unless the disease variant itself is genotyped.     

Prediction of haplotypes for ungenotyped animals and its effect on marker-assisted breeding value estimation

Background

In livestock populations, missing genotypes on a large proportion of animals are a major problem to implement the estimation of marker-assisted breeding values using haplotypes. The objective of this article is to develop a method to predict haplotypes of animals that are not genotyped using mixed model equations and to investigate the effect of using these predicted haplotypes on the accuracy of marker-assisted breeding value estimation.

Methods

For genotyped animals, haplotypes were determined and for each animal the number of haplotype copies (nhc) was counted, i.e. 0, 1 or 2 copies. In a mixed model framework, nhc for each haplotype were predicted for ungenotyped animals as well as for genotyped animals using the additive genetic relationship matrix. The heritability of nhc was assumed to be 0.99, allowing for minor genotyping and haplotyping errors. The predicted nhc were subsequently used in marker-assisted breeding value estimation by applying random regression on these covariables. To evaluate the method, a population was simulated with one additive QTL and an additive polygenic genetic effect. The QTL was located in the middle of a haplotype based on SNP-markers.

Results

The accuracy of predicted haplotype copies for ungenotyped animals ranged between 0.59 and 0.64 depending on haplotype length. Because powerful BLUP-software was used, the method was computationally very efficient. The accuracy of total EBV increased for genotyped animals when marker-assisted breeding value estimation was compared with conventional breeding value estimation, but for ungenotyped animals the increase was marginal unless the heritability was smaller than 0.1. Haplotypes based on four markers yielded the highest accuracies and when only the nearest left marker was used, it yielded the lowest accuracy. The accuracy increased with increasing marker density. Accuracy of the total EBV approached that of gene-assisted BLUP when 4-marker haplotypes were used with a distance of 0.1 cM between the markers.

Conclusions

The proposed method is computationally very efficient and suitable for marker-assisted breeding value estimation in large livestock populations including effects of a number of known QTL. Marker-assisted breeding value estimation using predicted haplotypes increases accuracy especially for traits with low heritability.     

Extending long-range phasing and haplotype library imputation methods to impute genotypes on sex chromosomes

AlphaImpute is a flexible and accurate genotype imputation tool that was originally designed for the imputation of genotypes on autosomal chromosomes. In some species, sex chromosomes comprise a large portion of the genome. For example, chromosome Z represents approximately 8% of the chicken genome and therefore is likely to be important in determining genetic variation in a population. When breeding programs make selection decisions based on genomic information, chromosomes that are not represented on the genotyping platform will not be subject to selection. Therefore imputation algorithms should be able to impute genotypes for all chromosomes. The objective of this research was to extend AlphaImpute so that it could impute genotypes on sex chromosomes. The accuracy of imputation was assessed using different genotyping strategies in a real commercial chicken population. The correlation between true and imputed genotypes was high in all the scenarios and was 0.96 for the most favourable scenario. Overall, the accuracy of imputation of the sex chromosome was slightly lower than that of autosomes for all scenarios considered.     

An exploration of sex-specific linkage disequilibrium on chromosome X in Caucasians from the COGA study

This paper explores the decay of linkage disequilibrium (LD) on the autosomes and chromosome X. The extent of marker-marker LD is important for both linkage and association studies. The analysis of the Caucasian sample from the Collaborative Study on the Genetics of Alcoholism study revealed the expected negative relationship between the magnitude of the marker-marker LD and distance (cM), with the male and female subgroups exhibiting similar patterns of LD. The observed extent of LD in females was less across the pseudoautosomal markers relative to the heterosomal region of chromosome X. Marked differences in LD patterns were also observed between chromosomes X and the 22 autosomes in both males and females.     

The extent of linkage disequilibrium in beef cattle breeds using high-density SNP genotypes

Background

The extent of linkage disequilibrium (LD) between molecular markers impacts genome-wide association studies and implementation of genomic selection. The availability of high-density single nucleotide polymorphism (SNP) genotyping platforms makes it possible to investigate LD at an unprecedented resolution. In this work, we characterised LD decay in breeds of beef cattle of taurine, indicine and composite origins and explored its variation across autosomes and the X chromosome.

Findings

In each breed, LD decayed rapidly and r² was less than 0.2 for marker pairs separated by 50 kb. The LD decay curves clustered into three groups of similar LD decay that distinguished the three main cattle types. At short distances between markers (< 10 kb), taurine breeds showed higher LD (r² = 0.45) than their indicine (r² = 0.25) and composite (r² = 0.32) counterparts. This higher LD in taurine breeds was attributed to a smaller effective population size and a stronger bottleneck during breed formation. Using all SNPs on only the X chromosome, the three cattle types could still be distinguished. However for taurine breeds, the LD decay on the X chromosome was much faster and the background level much lower than for indicine breeds and composite populations. When using only SNPs that were polymorphic in all breeds, the analysis of the X chromosome mimicked that of the autosomes.

Conclusions

The pattern of LD mirrored some aspects of the history of breed populations and showed a sharp decay with increasing physical distance between markers. We conclude that the availability of the HD chip can be used to detect association signals that remained hidden when using lower density genotyping platforms, since LD dropped below 0.2 at distances of 50 kb.     

Genomic relationships based on X chromosome markers and accuracy of genomic predictions with and without X chromosome markers

Background

Although the X chromosome is the second largest bovine chromosome, markers on the X chromosome are not used for genomic prediction in some countries and populations. In this study, we presented a method for computing genomic relationships using X chromosome markers, investigated the accuracy of imputation from a low density (7K) to the 54K SNP (single nucleotide polymorphism) panel, and compared the accuracy of genomic prediction with and without using X chromosome markers.

Methods

The impact of considering X chromosome markers on prediction accuracy was assessed using data from Nordic Holstein bulls and different sets of SNPs: (a) the 54K SNPs for reference and test animals, (b) SNPs imputed from the 7K to the 54K SNP panel for test animals, (c) SNPs imputed from the 7K to the 54K panel for half of the reference animals, and (d) the 7K SNP panel for all animals. Beagle and Findhap were used for imputation. GBLUP (genomic best linear unbiased prediction) models with or without X chromosome markers and with or without a residual polygenic effect were used to predict genomic breeding values for 15 traits.

Results

Averaged over the two imputation datasets, correlation coefficients between imputed and true genotypes for autosomal markers, pseudo-autosomal markers, and X-specific markers were 0.971, 0.831 and 0.935 when using Findhap, and 0.983, 0.856 and 0.937 when using Beagle. Estimated reliabilities of genomic predictions based on the imputed datasets using Findhap or Beagle were very close to those using the real 54K data. Genomic prediction using all markers gave slightly higher reliabilities than predictions without X chromosome markers. Based on our data which included only bulls, using a G matrix that accounted for sex-linked relationships did not improve prediction, compared with a G matrix that did not account for sex-linked relationships. A model that included a polygenic effect did not recover the loss of prediction accuracy from exclusion of X chromosome markers.

Conclusions

The results from this study suggest that markers on the X chromosome contribute to accuracy of genomic predictions and should be used for routine genomic evaluation.     

The effect of linkage disequilibrium on linkage analysis of incomplete pedigrees

Dense SNP maps can be highly informative for linkage studies. But when parental genotypes are missing, multipoint linkage scores can be inflated in regions with substantial marker-marker linkage disequilibrium (LD). Such regions were observed in the Affymetrix SNP genotypes for the Genetic Analysis Workshop 14 (GAW14) Collaborative Study on the Genetics of Alcoholism (COGA) dataset, providing an opportunity to test a novel simulation strategy for studying this problem. First, an inheritance vector (with or without linkage present) is simulated for each replicate, i.e., locations of recombinations and transmission of parental chromosomes are determined for each meiosis. Then, two sets of founder haplotypes are superimposed onto the inheritance vector: one set that is inferred from the actual data and which contains the pattern of LD; and one set created by randomly selecting parental alleles based on the known allele frequencies, with no correlation (LD) between markers. Applying this strategy to a map of 176 SNPs (66 Mb of chromosome 7) for 100 replicates of 116 sibling pairs, significant inflation of multipoint linkage scores was observed in regions of high LD when parental genotypes were set to missing, with no linkage present. Similar inflation was observed in analyses of the COGA data for these affected sib pairs with parental genotypes set to missing, but not after reducing the marker map until r2 between any pair of markers was 相似文献   

A new method for haplotype inference including full-sib information

Recent literature has suggested that haplotype inference through close relatives, especially from nuclear families, can be an alternative strategy in determining linkage phase and estimating haplotype frequencies. In the case of no possibility to obtain genotypes for parents, and only full-sib information being used, a new approach is suggested to infer phase and to reconstruct haplotypes. We present a maximum-likelihood method via an expectation-maximization algorithm, called FSHAP, using only full-sib information when parent information is not available. FSHAP can deal with families with an arbitrary number of children, and missing parents or missing genotypes can be handled as well. In a simulation study we compare FSHAP with another existing expectation-maximization (EM)-based approach (FAMHAP), the conditioning approach implemented in FBAT and GENEHUNTER, which is only pedigree based and assumes linkage equilibrium. In most situations, FSHAP has the smallest discrepancy of haplotype frequency estimation and the lowest error rate in haplotype reconstruction, only in some cases FAMHAP yields comparable results. GENEHUNTER produces the largest discrepancy, and FBAT produces the highest error rate in offspring in most situations. Among the methods compared, FSHAP has the highest accuracy in reconstructing the diplotypes of the unavailable parents. Potential limitations of the method, e.g., in analyzing very large haplotypes, are indicated and possible solutions are discussed.     

The effect of missing marker genotypes on the accuracy of gene-assisted breeding value estimation: a comparison of methods

In livestock populations, missing genotypes on a large proportion of the animals is a major problem when implementing gene-assisted breeding value estimation for genes with known effect. The objective of this study was to compare different methods to deal with missing genotypes on accuracy of gene-assisted breeding value estimation for identified bi-allelic genes using Monte Carlo simulation. A nested full-sib half-sib structure was simulated with a mixed inheritance model with one bi-allelic quantitative trait loci (QTL) and a polygenic effect due to infinite number of polygenes. The effect of the QTL was included in gene-assisted BLUP either by random regression on predicted gene content, i.e. the number of positive alleles, or including haplotype effects in the model with an inverse IBD matrix to account for identity-by-descent relationships between haplotypes using linkage analysis information (IBD-LA). The inverse IBD matrix was constructed using segregation indicator probabilities obtained from multiple marker iterative peeling. Gene contents for unknown genotypes were predicted using either multiple marker iterative peeling or mixed model methodology. For both methods, gene-assisted breeding value estimation increased accuracies of total estimated breeding value (EBV) with 0% to 22% for genotyped animals in comparison to conventional breeding value estimation. For animals that were not genotyped, the increase in accuracy was much lower (0% to 5%), but still substantial when the heritability was 0.1 and when the QTL explained at least 15% of the genetic variance. Regression on predicted gene content yielded higher accuracies than IBD-LA. Allele substitution effects were, however, overestimated, especially when only sires and males in the last generation were genotyped. For juveniles without phenotypic records and traits measured only on females, the superiority of regression on gene content over IBD-LA was larger than when all animals had phenotypes. Missing gene contents were predicted with higher accuracy using multiple-marker iterative peeling than with using mixed model methodology, but the difference in accuracy of total EBV was negligible and mixed model methodology was computationally much faster than multiple iterative peeling. For large livestock populations it can be concluded that gene-assisted breeding value estimation can be practically best performed by regression on gene contents, using mixed model methodology to predict missing marker genotypes, combining phenotypic information of genotyped and ungenotyped animals in one evaluation. This technique would be, in principle, also feasible for genomic selection. It is expected that genomic selection for ungenotyped animals using predicted single nucleotide polymorphism gene contents might be beneficial especially for low heritable traits.     

Sexual antagonism and meiotic drive cause stable linkage disequilibrium and favour reduced recombination on the X chromosome

Sexual antagonism and meiotic drive are sex‐specific evolutionary forces with the potential to shape genomic architecture. Previous theory has found that pairing two sexually antagonistic loci or combining sexual antagonism with meiotic drive at linked autosomal loci augments genetic variation, produces stable linkage disequilibrium (LD) and favours reduced recombination. However, the influence of these two forces has not been examined on the X chromosome, which is thought to be enriched for sexual antagonism and meiotic drive. We investigate the evolution of the X chromosome under both sexual antagonism and meiotic drive with two models: in one, both loci experience sexual antagonism; in the other, we pair a meiotic drive locus with a sexually antagonistic locus. We find that LD arises between the two loci in both models, even when the two loci freely recombine in females and that driving haplotypes will be enriched for male‐beneficial alleles, further skewing sex ratios in these populations. We introduce a new measure of LD, , which accounts for population allele frequencies and is appropriate for instances where these are sex specific. Both models demonstrate that natural selection favours modifiers that reduce the recombination rate. These results inform observed patterns of congealment found on driving X chromosomes and have implications for patterns of natural variation and the evolution of recombination rates on the X chromosome.     

2LD, GENECOUNTING and HAP: Computer programs for linkage disequilibrium analysis

Computer programs are introduced which calculate pair-wise linkage disequilibrium statistics and conduct haplotype frequency estimation, including X chromosome data, and using a heuristic algorithm to handle multiple genetic markers and missing data. AVAILABILITY: Programs 2LD, GENECOUNTING and HAP are available on Internet from http://www.hgmp.mrc.ac.uk/~jzhao and http://www.iop.kcl.ac.uk/IoP/Departments/PsychMed/GEpiBSt/software.shtml