The Value of Molecular Haplotypes in a Family-Based Linkage Study
2006; Elsevier BV; Volume: 79; Issue: 3 Linguagem: Inglês
10.1086/506626
ISSN1537-6605
AutoresElizabeth M. Gillanders, John V. Pearson, Alexa J.M. Sorant, Jeffrey M. Trent, Jeff O’Connell, Joan E. Bailey‐Wilson,
Tópico(s)Digestive system and related health
ResumoNovel methods that could improve the power of conventional methods of gene discovery for complex diseases should be investigated. In a simulation study, we aimed to investigate the value of molecular haplotypes in the context of a family-based linkage study. The term “haplotype” (or “haploid genotype”) refers to syntenic alleles inherited on a single chromosome, and we use the term “molecular haplotype” to refer to haplotypes that have been determined directly by use of a molecular technique such as long-range allele-specific polymerase chain reaction. In our study, we simulated genotype and phenotype data and then compared the powers of analyzing these data under the assumptions that various levels of information from molecular haplotypes were available. (This information was available because of the simulation procedure.) Several conclusions can be drawn. First, as expected, when genetic homogeneity is expected or when marker data are complete, it is not efficient to generate molecular haplotyping information. However, with levels of heterogeneity and missing data patterns typical of complex diseases, we observed a 23%–77% relative increase in the power to detect linkage in the presence of heterogeneity with heterogeneity LOD scores >3.0 when all individuals are molecularly haplotyped (compared with the power when only standard genotypes are used). Furthermore, our simulations indicate that most of the increase in power can be achieved by molecularly haplotyping a single individual in each family, thereby making molecular haplotyping a valuable strategy for increasing the power of gene mapping studies of complex diseases. Maximization of power, given an existing family set, can be particularly important for late-onset, often-fatal diseases such as cancer, for which informative families are difficult to collect. Novel methods that could improve the power of conventional methods of gene discovery for complex diseases should be investigated. In a simulation study, we aimed to investigate the value of molecular haplotypes in the context of a family-based linkage study. The term “haplotype” (or “haploid genotype”) refers to syntenic alleles inherited on a single chromosome, and we use the term “molecular haplotype” to refer to haplotypes that have been determined directly by use of a molecular technique such as long-range allele-specific polymerase chain reaction. In our study, we simulated genotype and phenotype data and then compared the powers of analyzing these data under the assumptions that various levels of information from molecular haplotypes were available. (This information was available because of the simulation procedure.) Several conclusions can be drawn. First, as expected, when genetic homogeneity is expected or when marker data are complete, it is not efficient to generate molecular haplotyping information. However, with levels of heterogeneity and missing data patterns typical of complex diseases, we observed a 23%–77% relative increase in the power to detect linkage in the presence of heterogeneity with heterogeneity LOD scores >3.0 when all individuals are molecularly haplotyped (compared with the power when only standard genotypes are used). Furthermore, our simulations indicate that most of the increase in power can be achieved by molecularly haplotyping a single individual in each family, thereby making molecular haplotyping a valuable strategy for increasing the power of gene mapping studies of complex diseases. Maximization of power, given an existing family set, can be particularly important for late-onset, often-fatal diseases such as cancer, for which informative families are difficult to collect. In contrast to simple Mendelian disorders, susceptibility to common complex diseases such as cancer, type-2 diabetes, or Alzheimer disease is multifactorial and involves multiple genetic and environmental risk factors. Efforts to localize susceptibility genes involved in complex diseases have been limited by genetic heterogeneity, incomplete penetrance, phenocopies, and, frequently, late age at disease onset. Each of these factors can result in a significant reduction in statistical power for any individual gene-mapping study. Thus, novel methods that could improve the power of traditional methods of gene discovery for complex diseases should be examined. Complex diseases include most common diseases of adult life, and they account for most human morbidity and mortality. Therefore, improvements in the methods used to decipher their genetic etiologies should be of paramount importance. Reconstruction of haplotypes has proven critical to several studies that have succeeded in identifying genetic factors involved in complex-trait susceptibility.1Puffenberger EG Kauffman ER Bolk S Matise TC Washington SS Angrist M Weissenbach J Garver KL Mascari M Ladda R Siaugenhaupt S Chakravarti A Identity-by-descent and association mapping of a recessive gene for Hirschsprung disease on human chromosome 13q22.Hum Mol Genet. 1994; 3: 1217-1225Crossref PubMed Scopus (209) Google Scholar, 2Drysdale CM McGraw DW Stack CB Stephens JC Judson RS Nandabalan K Arnold K Ruano G Liggett SB Complex promoter and coding region β 2-adrenergic receptor haplotypes alter receptor expression and predict in vivo responsiveness.Proc Natl Acad Sci USA. 2000; 97: 10483-10488Crossref PubMed Scopus (893) Google Scholar, 3Hugot JP Chamaillard M Zouali H Lesage S Cezard JP Belaiche J Almer S Tysk C O’Morain CA Gassull M Binder V Finkel Y Cortot A Modigliani R Laurent-Puig P Gower-Rousseau C Macry J Colombel JF Sahbatou M Thomas G Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn’s disease.Nature. 2001; 411: 599-603Crossref PubMed Scopus (4674) Google Scholar, 4Rioux JD Daly MJ Silverberg MS Lindblad K Steinhart H Cohen Z Delmonte T et al.Genetic variation in the 5q31 cytokine gene cluster confers susceptibility to Crohn disease.Nat Genet. 2001; 29: 223-228Crossref PubMed Scopus (693) Google Scholar Haplotypes provide additional information to both linkage and linkage disequilibrium (LD) studies and therefore may facilitate the mapping of a disease gene by allowing a more precise localization of the gene within a chromosomal region initially identified by linkage analysis. Candidate regions for complex diseases, initially identified by genomewide linkage scans, can often be prohibitively large (20–30 cM). These regions can contain upwards of 100 genes, which requires further narrowing of the candidate interval before positional cloning efforts. Recent advances in molecular technologies and the availability of the human genome sequence have revolutionized researchers’ ability to catalogue human genetic variation. However, reconstruction of haplotypes from conventional genotypes in diploid organisms such as humans can be complicated, since the parental origins of the two alleles of each genotype are not directly observable. There are three principle haplotyping approaches: (1) statistical estimation, (2) inference from family data, and (3) empirical (or “direct”) molecular haplotyping. The reliability of statistical methods in reconstruction of haplotypes depends on the number of markers, allele frequencies, fraction of missing data, genotyping error rate, and LD between markers.5Fallin D Schork NJ Accuracy of haplotype frequency estimation for biallelic loci, via the expectation-maximization algorithm for unphased diploid genotype data.Am J Hum Genet. 2000; 67: 947-959Abstract Full Text Full Text PDF PubMed Scopus (337) Google Scholar, 6Tishkoff SA Pakstis AJ Ruano G Kidd KK The accuracy of statistical methods for estimation of haplotype frequencies: an example from the CD4 locus.Am J Hum Genet. 2000; 67: 518-522Abstract Full Text Full Text PDF PubMed Scopus (106) Google Scholar, 7Kirk KM Cardon LR The impact of genotyping error on haplotype reconstruction and frequency estimation.Eur J Hum Genet. 2002; 10: 616-622Crossref PubMed Scopus (67) Google Scholar Inferring phase from family data can be limited by uninformative or missing genotypes. In addition, late age at onset for many complex diseases can preclude collection of DNA samples from previous generations, thereby further limiting strategies to reconstruct haplotypes with use of family data. In contrast, molecular haplotyping methods are empirical and are not dependent on statistical assumptions or estimation. Two popular molecular haplotyping methods include (1) long-range, allele-specific PCR (AS-PCR)8Michalatos-Beloin S Tishkoff SA Bentley KL Kidd KK Ruano G Molecular haplotyping of genetic markers 10 kb apart by allele-specific long-range PCR.Nucleic Acids Res. 1996; 24: 4841-4843Crossref PubMed Scopus (101) Google Scholar, 9Antonellis A Rogus JJ Canani LH Makita Y Pezzolesi MG Nam M Ng D Moczulski D Warram JH Krolewski AS A method for developing high-density SNP maps and its application at the type 1 angiotensin II receptor (AGTR1) locus.Genomics. 2002; 79: 326-332Crossref PubMed Scopus (15) Google Scholar, 10Yu CE Devlin B Galloway N Loomis E Schellenberg GD ADLAPH: a molecular haplotyping method based on allele-discriminating long-range PCR.Genomics. 2004; 84: 600-612Crossref PubMed Scopus (16) Google Scholar and (2) diploid-to-haploid methods, such as conversion.11Papadopoulos N Leach FS Kinzler KW Vogelstein B Monoallelic mutation analysis (MAMA) for identifying germline mutations.Nat Genet. 1995; 11: 99-102Crossref PubMed Scopus (50) Google Scholar Crawford and Nickerson12Crawford DC Nickerson DA Definition and clinical importance of haplotypes.Annu Rev Med. 2005; 56: 303-320Crossref PubMed Scopus (262) Google Scholar describe these methods in detail. In brief, these molecular techniques unambiguously reconstruct haplotypes in the following manner. AS-PCR involves selective PCR amplification of one of the two chromosomes at a given heterozygous locus. This is frequently done by designing PCR primers that will match (or mismatch) one allele at the 3′ end of the primer. By use of long-range PCR methods, a molecular haplotype of up to 40 kb can be determined. The conversion method entails generation of mouse-human somatic cell hybrids, which retain only a subset of human chromosomes. Once hybrids that are monosomic for the chromosome of interest are identified, haplotypes can be reconstructed with conventional genotyping of the haploid cells. In short, both methods provide unequivocal molecular haplotypes. Several studies have provided evidence of the value of molecular haplotyping in the context of LD studies.13Douglas JA Boehnke M Gillanders E Trent JM Gruber SB Experimentally-derived haplotypes substantially increase the efficiency of linkage disequilibrium studies.Nat Genet. 2001; 28: 361-364Crossref PubMed Scopus (134) Google Scholar, 14Schaid DJ Relative efficiency of ambiguous vs directly measured haplotype frequencies.Genet Epidemiol. 2002; 23: 426-443Crossref PubMed Scopus (59) Google Scholar, 15Thomas S Porteous D Visscher PM Power of direct vs indirect haplotyping in association studies.Genet Epidemiol. 2004; 26: 116-124Crossref PubMed Scopus (10) Google Scholar In this study, we used simulations to compare the power of using various levels of molecular haplotypes (available because of the simulation procedure) with the power of using standard genotyping in the context of a family-based linkage study. To clarify, we are not assuming any molecular haplotyping method in particular; we are simply assuming that we have molecularly derived haplotypes available. These results will need to be considered within the context of considerably increased laboratory expenditure. Current molecular haplotyping methods are fairly limited and not particularly well suited for high-throughput work; however, positive results might motivate new molecular or statistical approaches that could more easily capitalize on the benefits of molecular haplotyping information. Using the Genometric Analysis Simulation Program (GASP),16Wilson AF Bailey-Wilson JE Pugh EW Sorant AJM The Genometric Analysis Simulation Program (GASP): a software tool for testing and investigating methods in statistical genetics.Am J Hum Genet. 1996; 59: A193Google Scholar we simulated qualitative trait and marker data for the three-generation pedigree structure shown in figure 1. The simulated qualitative trait was due to a single locus with an autosomal dominant mode of inheritance and a disease allele (D) frequency equal to 0.01. Individuals with both the DD and Dd genotypes had an 80% probability of developing the disease. Of individuals with a normal genotype (dd), 4% developed the trait. All individuals within the pedigree were considered to be beyond the age of risk. Families were ascertained (i.e., selected from randomly simulated families to be in the analyzed sample) on the basis of having a minimum of three affected members who were at least minimally informative for linkage analysis. Specifically, families were excluded if the three affected members were (1) all founders (fig. 2A), (2) a parent-parent-offspring trio (fig. 2B), (3) a parent-offspring pair and a founder who is not a grandparent (fig. 2C), (4) a founder-founder-spouse trio (fig. 2D), or (5) a trio comprising individuals 7, 12, and either 5 or 6 (fig. 2E). This ascertainment procedure was designed to mimic the real procedures used in linkage studies of qualitative traits. Simulations of families for each replicate continued until there were 100 pedigrees that met our ascertainment criteria. Examples of pedigrees included in our simulation are shown in figure 3.Figure 3Examples of specific pedigrees included in simulation analysesView Large Image Figure ViewerDownload Hi-res image Download (PPT) GASP was used to simulate genotype data for eight STR markers, including four STR genomewide-scan (GWS) markers that were 10 cM apart and four STR fine-mapping (FM) markers that were 1 cM apart. Each simulated STR marker had five equally frequent alleles. In separate simulations, we generated genotype data for four SNP markers. These SNP GWS markers were 1 cM apart and had a minor-allele frequency of 0.40. The genetic marker loci were all assumed to be in linkage equilibrium. Furthermore, because the marker data were simulated, the haplotypes were known with certainty for all individuals in the data set. Therefore, unlike a real linkage study (where only the genotype at each marker is known), we were able to write out the simulated marker data showing which allele was maternally or paternally derived for all loci in all individuals. This is the information that would be available if that person were molecularly haplotyped. We then included various levels of these simulated molecular haplotypes in the linkage analyses, as described below. All simulations are outlined in figures Figure 4, Figure 5, Figure 6, Figure 7. In each simulated family, the trait could be either linked or unlinked to the analyzed marker set. We simulated three different levels of genetic heterogeneity in which the percentage of linked families varied: (1) none of the families linked to the marker set (H0), (2) 100% of the families linked to the marker set (H1), and (3) 25% of the families linked to the marker set (H2). The H0 level of heterogeneity was simulated, to confirm that the inclusion of molecular haplotypes did not cause increases in type I error rates. The H1 and H2 levels of heterogeneity were used to determine the extent to which the inclusion of molecular haplotype information into the linkage analysis improved power.Figure 5Outline of additional STR FM analyses (H2 only). For the H2 level of heterogeneity only (A), we considered two additional levels of molecular haplotyping information (B), each with molecular haplotyping information available for a single individual. For each of these two levels of molecular haplotyping information, we considered three levels of missing data (M0, M1, and M2) (C). Results are shown in table 4.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 6Outline of additional STR GWS analyses (H2 only). For the H2 level of heterogeneity only, we also analyzed four STR markers at GWS density (markers 10 cM apart). Results are shown in table 5.View Large Image Figure ViewerDownload Hi-res image Download (PPT)Figure 7Outline of additional SNP GWS analyses (H2 only). For the H2 level of heterogeneity only, we also analyzed four SNP markers at GWS density (markers 1 cM apart). Results are shown in table 6.View Large Image Figure ViewerDownload Hi-res image Download (PPT) For each level of heterogeneity (H0, H1, and H2), at least nine different models of data availability were considered (fig. 4). Specifically, for each replicate, we considered three different levels of haplotyping information: (1) no molecular haplotypes were included in the linkage analysis (only individual genotypes were used) (P0), (2) simulated molecular haplotypes of affected individuals were included in the linkage analysis (only genotypes of unaffected individuals were used) (P1), and (3) simulated molecular haplotypes from all genotyped individuals were included in the linkage analysis (P2). These different levels of simulated molecular haplotyping were used to determine whether all genotyped persons in the pedigree needed to be molecularly haplotyped to increase linkage power or whether linkage power gains could be obtained by performing molecular haplotyping on a smaller proportion of the family members. For each level of molecular haplotyping information (P0, P1, and P2), we considered three levels of missing marker data (i.e., different patterns of ungenotyped family members): (1) no missing data (marker data available for everyone) (M0), (2) marker data missing for individuals in our top generation of our simulated pedigree (individuals 1 and 2) (M1), and (3) marker data missing for individuals in our top generation as well as for 50% of individuals in our second generation (a randomly chosen half of founders and a randomly chosen half of nonfounders) (M2). We used these different missing data rates to evaluate whether including molecular haplotype data in the linkage analysis always increased power or whether it increased power only in the presence of incomplete family genotype data. When marker data were missing for an individual, that person’s genotypes (at all marker loci) and simulated molecular haplotype were all treated as unknown in the linkage analysis. For each of these 27 models (nine basic data availability combinations times three levels of genetic heterogeneity), a set of four STR FM (1-cM spacing) markers surrounding the trait locus was analyzed (fig. 4). For heterogeneity model H2 only, these four FM STR markers were also analyzed for two more levels of haplotyping information: simulated molecular haplotyping for one randomly chosen member of our middle generation (P3) and simulated molecular haplotyping for one randomly chosen member of our bottom generation (P4), for each level of missing data (M1, M2, and M3) (fig. 5). Once again, these two levels of simulated molecular haplotyping were used to determine whether all genotyped persons in the pedigree needed to be molecularly haplotyped to increase linkage power or whether linkage power gains could be obtained by performing molecular haplotyping on only a small proportion of the family members. Again, for heterogeneity model H2 only and the basic nine data-availability combinations, we also analyzed two additional marker sets: four GWS STR markers 10 cM apart (fig. 6) and four GWS SNP markers 1 cM apart (fig. 7). For each described scenario, we analyzed 1,000 replicates. We used version 2 of the VITESSE program17O’Connell JR Weeks DE The VITESSE algorithm for rapid exact multilocus linkage analysis via genotype set-recoding and fuzzy inheritance.Nat Genet. 1995; 11: 402-408Crossref PubMed Scopus (423) Google Scholar, 18O’Connell JR Rapid multipoint linkage analysis via inheritance vectors in the Elston-Stewart algorithm.Hum Hered. 2001; 51: 226-240Crossref PubMed Scopus (32) Google Scholar to perform multipoint linkage analyses. VITESSE is a linkage-analysis program that uses the Elston-Stewart algorithm to compute the likelihood of pedigree data. VITESSE uses a novel set-recoding scheme to reduce the number of genotypes required in the likelihood calculation, thereby improving the computational performance. The program accepts a wide variety of special input formats, including phased genotype data. Standard genotype-input format does not distinguish the parental source of alleles in heterozygous genotypes. VITESSE permits a vertical bar delimiter (|) between alleles to set the paternal source of the allele. For example, an input genotype of 1|2 in the pedigree file is read by VITESSE as specifying the alleles 1 and 2 as paternally and maternally inherited, respectively. Moreover, this input format can be used to specify paternally inherited n-locus haplotypes at n loci by specifying phased genotypes at each of the n loci. This is how our simulated molecular haplotypes were specified in the linkage analysis. Multipoint linkage analyses were performed under the assumption that the genetic model is the same as the generating one. In multipoint likelihood-based linkage analysis, the likelihood of the marker data conceptually is a sum over all possible haplotype configurations, and phase information for one or more individuals in the family reduces the number of configurations. When VITESSE reads in molecularly determined haplotypes as described above, a reduction in the number of possible haplotype configurations in the family is done at the time of genotype input. Thereafter, VITESSE performs the standard likelihood calculations over the remaining possible haplotype configurations in the family. It is this reduction in the uncertainty about the true haplotype configuration in the family that is expected to increase power to detect linkage when molecular haplotypes are included in the linkage analysis. Heterogeneity LOD–score (HLOD) calculations to test for linkage in the presence of genetic heterogeneity were performed using the admixture test.19Ott J Analysis of human genetic linkage. rev ed. Johns Hopkins University Press, Baltimore1991Google Scholar Power to detect linkage (or, in the completely unlinked situation, type I error rate) was measured as the percentage of 1,000 replicates that reached an HLOD of at least 1.0, at least 2.0, and at least 3.0. Analyses were run on the TGen Research Institute IBM 1350 computational cluster. The cluster had 512 IBM X Series computational nodes, each with two 2.4 GHz Intel Xeon processors, for a total of 1,024 central processing units. Each node had 2 GB of RAM and Gigabit Ethernet network connections. The operating system was RedHat Enterprise Linux 3.0, PBSPro was used to handle job scheduling, and the IBM CSM (Cluster System Management) was used to monitor and maintain the cluster. Of the nodes, 128 are equipped with low-latency, high-throughput Myrinet interconnects. These nodes also have an extra 2 GB of RAM per node to allow for memory-intensive computations. PBSPro manages these resources and allows jobs to be run exclusively on these Myrinet nodes. All cluster nodes have access to a shared parallel 1 TB file system (IBM GPFS), which allows each individual node to read and write to the same data files simultaneously across the cluster. The GPFS file system uses IBM FAStT SAN (storage area network) storage units, providing high performance and high reliability. Summary results for the four STR FM marker multipoint analyses are presented in tables 1, 2, and 3 for genetic heterogeneity models H0, H1, and H2, respectively. In all simulations, models 1–3 assume no one has been molecularly haplotyped (P0); thus, only genotype data were included in the linkage analysis. Models 4–6 were analyzed under the assumption that all genotyped affected individuals had been molecularly haplotyped (P1) (with use of some molecular method), so that simulated molecular haplotypes for the affected individuals as well as genotypes for unaffected individuals were used in the analysis. Finally, models 7–9 assume everyone has been molecularly haplotyped (P2), so that simulated molecular haplotypes were included in the analysis of all genotyped family members. In table 4 (heterogeneity model H2 only), haplotyping schemes P3 and P4 (one family member molecularly haplotyped) are presented. In these models (10–15) the simulated molecular haplotype for one individual and genotypes for the remaining family members were used in the analysis. For each phase information level, three missing marker data options (M0, M1, and M2) were considered.Table 1Type I Error of Four STR FM Markers (1 cM Apart), with 0% Families Linked (H0)Percentage of 1,000 Replicates with HLODModelMolecular Haplotyping InformationOptions for Missing Marker Data>1.0>2.0>3.01All members genotyped (P0)No marker data missing (M0)1.8.2.02All members genotyped (P0)Marker data missing for top generation (M1)1.6.1.03All members genotyped (P0)Marker data missing for top generation and 50% of middle generation (M2)1.9.2.04Affected individuals haplotyped (P1)No marker data missing (M0)1.8.2.05Affected individuals haplotyped (P1)Marker data missing for top generation (M1)1.8.3.06Affected individuals haplotyped (P1)Marker data missing for top generation and 50% of middle generation (M2)1.2.2.07All members haplotyped (P2)No marker data missing (M0)1.8.2.08All members haplotyped (P2)Marker data missing for top generation (M1)1.8.3.09All members haplotyped (P2)Marker data missing for top generation and 50% of middle generation (M2)1.9.1.1 Open table in a new tab Table 2Power of Four STR FM Markers (1 cM Apart), with 100% of Families Linked (H1)Percentage of 1,000 Replicates with HLODModelMolecular Haplotyping InformationOptions for Missing Marker Data>1.0>2.0>3.01All members genotyped (P0)No marker data missing (M0)100.0100.0100.02All members genotyped (P0)Marker data missing for top generation (M1)100.0100.0100.03All members genotyped (P0)Marker data missing for top generation and 50% of middle generation (M2)100.0100.0100.04Affected individuals haplotyped (P1)No marker data missing (M0)100.0100.0100.05Affected individuals haplotyped (P1)Marker data missing for top generation (M1)100.0100.0100.06Affected individuals haplotyped (P1)Marker data missing for top generation and 50% of middle generation (M2)100.0100.0100.07All members haplotyped (P2)No marker data missing (M0)100.0100.0100.08All members haplotyped (P2)Marker data missing for top generation (M1)100.0100.0100.09All members haplotyped (P2)Marker data missing for top generation and 50% of middle generation (M2)100.0100.0100.0 Open table in a new tab Table 3Power of Four STR FM Markers (1 cM Apart), with 25% of Families Linked (H2)Percentage of 1,000 Replicates with HLODModelMolecular Haplotyping InformationOptions for Missing Marker Data>1.0>2.0>3.01All members genotyped (P0)No marker data missing (M0)98.694.587.02All members genotyped (P0)Marker data missing for top generation (M1)97.388.874.93All members genotyped (P0)Marker data missing for top generation and 50% of middle generation (M2)95.081.063.14Affected individuals haplotyped (P1)No marker data missing (M0)98.694.587.05Affected individuals haplotyped (P1)Marker data missing for top generation (M1)98.794.386.66Affected individuals haplotyped (P1)Marker data missing for top generation and 50% of middle generation (M2)97.589.374.47All members haplotyped (P2)No marker data missing (M0)98.694.587.08All members haplotyped (P2)Marker data missing for top generation (M1)98.794.687.09All members haplotyped (P2)Marker data missing for top generation and 50% of middle generation (M2)98.190.477.4 Open table in a new tab Table 4Power of Four STR FM Markers (1 cM Apart), with 25% of Families Linked (H2) and with Additional Molecular Haplotyping LevelsPercentage of 1,000 Replicates with HLODModelMolecular Haplotyping InformationOptions for Missing Marker Data>1.0>2.0>3.010One individual haplotyped (middle generation) (P3)No marker data missing (M0)98.694.587.011One individual haplotyped (middle generation) (P3)Marker data missing for top generation (M1)98.592.182.612One individual haplotyped (middle generation) (P3)Marker data missing for top generation and 50% of middle generation (M2)97.087.171.713One individual haplotyped (bottom generation) (P4)No marker data missing (M0)98.694.587.014One individual haplotyped (bottom generation) (P4)Marker data missing for top generation (M1)97.388.874.915One individual haplotyped (bottom generation) (P4)Marker data missing for top generation and 50% of middle generation (M2)94.781.162.8 Open table in a new tab Table 1 presents a summary of HLOD results for the replicates in which none of the simulated families have their trait locus linked to the markers tested (H0). In this situation, any linkage detected was a type I error. The HLOD results across different molecular ha
Referência(s)