Capacitating epistasis--detection and role in the genetic architecture of complex traits

Here, we discuss the potential role of capacitating epistasis in the genetic architecture of complex traits. Two alternative methods for identifying such gene-gene interactions in genetic association studies-mapping of variance controlling loci and the variance plane ratio (VPR) method-are introduced. An overview of the theoretical foundation of the methods is presented together with a discussion on their implementation and available software for performing these analyses. We conclude by highlighting a few examples of capacitating epistasis described in the literature and its potential impacts on the genetics of complex traits.

Identification of candidate genes and mutations in QTL regions for chicken growth using bioinformatic analysis of NGS and SNP-chip data

Mapping of chromosomal regions harboring genetic polymorphisms that regulate complex traits is usually followed by a search for the causative mutations underlying the observed effects. This is often a challenging task even after fine mapping, as millions of base pairs including many genes will typically need to be investigated. Thus to trace the causative mutation(s) there is a great need for efficient bioinformatic strategies. Here, we searched for genes and mutations regulating growth in the Virginia chicken lines - an experimental population comprising two lines that have been divergently selected for body weight at 56 days for more than 50 generations. Several quantitative trait loci (QTL) have been mapped in an F2 intercross between the lines, and the regions have subsequently been replicated and fine mapped using an Advanced Intercross Line. We have further analyzed the QTL regions where the largest genetic divergence between the High-Weight selected (HWS) and Low-Weight selected (LWS) lines was observed. Such regions, covering about 37% of the actual QTL regions, were identified by comparing the allele frequencies of the HWS and LWS lines using both individual 60K SNP chip genotyping of birds and analysis of read proportions from genome resequencing of DNA pools. Based on a combination of criteria including significance of the QTL, allele frequency difference of identified mutations between the selected lines, gene information on relevance for growth, and the predicted functional effects of identified mutations we propose here a subset of candidate mutations of highest priority for further evaluation in functional studies. The candidate mutations were identified within the GCG, IGFBP2, GRB14, CRIM1, FGF16, VEGFR-2, ALG11, EDN1, SNX6, and BIRC7 genes. We believe that the proposed method of combining different types of genomic information increases the probability that the genes underlying the observed QTL effects are represented among the candidate mutations identified.

Genetic dissection of growth traits in a Chinese indigenous × commercial broiler chicken cross

Background

In China, consumers often prefer indigenous broiler chickens over commercial breeds, as they have characteristic meat qualities requested within traditional culinary customs. However, the growth-rate of these indigenous breeds is slower than that of the commercial broilers, which means they have not yet reached their full economic value. Therefore, combining the valuable meat quality of the native chickens with the efficiency of the commercial broilers is of interest. In this study, we generated an F₂ intercross between the slow growing native broiler breed, Huiyang Beard chicken, and the fast growing commercial broiler breed, High Quality chicken Line A, and used it to map loci explaining the difference in growth rate between these breeds.

Results

A genome scan to identify main-effect loci affecting 24 growth-related traits revealed nine distinct QTL on six chromosomes. Many QTL were pleiotropic and conformed to the correlation patterns observed between phenotypes. Most of the mapped QTL were found in locations where growth QTL have been reported in other populations, although the effects were greater in this population. A genome scan for pairs of interacting loci identified a number of additional QTL in 10 other genomic regions. The epistatic pairs explained 6–8% of the residual phenotypic variance. Seven of the 10 epistatic QTL mapped in regions containing candidate genes in the ubiquitin mediated proteolysis pathway, suggesting the importance of this pathway in the regulation of growth in this chicken population.

Conclusions

The main-effect QTL detected using a standard one-dimensional genome scan accounted for a significant fraction of the observed phenotypic variance in this population. Furthermore, genes in known pathways present interesting candidates for further exploration. This study has thus located several QTL regions as promising candidates for further study, which will increase our understanding of the genetic mechanisms underlying growth-related traits in chickens.

PASE: a novel method for functional prediction of amino acid substitutions based on physicochemical properties

Background: Non-synonymous single-nucleotide polymorphisms (nsSNPs) within the coding regions of genes causing amino acid substitutions (AASs) may have a large impact on protein function. The possibilities to identify nsSNPs across genomes have increased notably with the advent of next-generation sequencing technologies. Thus, there is a strong need for efficient bioinformatics tools to predict the functional effect of AASs. Such tools can be used to identify the most promising candidate mutations for further experimental validation.

Results: Here we present prediction of AAS effects (PASE), a novel method that predicts the effect of an AASs based on physicochemical property changes. Evaluation of PASE, using a few AASs of known phenotypic effects and 3338 human AASs, for which functional effects have previously been scored with the widely used SIFT and PolyPhen tools, show that PASE is a useful method for functional prediction of AASs. We also show that the predictions can be further improved by combining PASE with information about evolutionary conservation.

Conclusion: PASE is a novel algorithm for predicting functional effects of AASs, which can be used for pinpointing the most interesting candidate mutations. PASE predictions are based on changes in seven physicochemical properties and can improve predictions from many other available tools, which are based on evolutionary conservation. Using available experimental data and predictions from the already existing tools, we demonstrate that PASE is a useful method for predicting functional effects of AASs, even when a limited number of query sequence homologs/orthologs are available.

Selection on variance-controlling genes: adaptability or stability

Simulations on a model system where a variance-controlling master locus scales the effects of a set of effector loci show that selection affects the variance-controlling locus more strongly than the effector loci, and that the direction of selection is dependent on the frequency of environmental changes.

Generation of a multi-locus chicken introgression line to study the effects of genetic interactions on metabolic phenotypes in chickens

Most biological traits are regulated by a complex interplay between genetic and environmental factors. By intercrossing divergent lines, it is possible to identify individual and interacting QTL involved in the genetic architecture of these traits. When the loci have been mapped, alternative strategies are needed for fine-mapping and studying the individual and interactive effects of the QTL in detail. We have previously identified, replicated, and fine mapped a four-locus QTL network that determines nearly half of the eightfold difference in body weight at 56 days of age between two divergently selected chicken lines. Here, we describe, to our knowledge, the first generation of a three-locus QTL introgression line in chickens. Recurrent marker-assisted backcrossing was used to simultaneously transfer QTL alleles from the low-weight selected line into the high-weight selected line. Three generations of backcrossing and one generation of intercrossing resulted in an introgression line where all three introgressed QTL and several unlinked and linked control-loci were segregating at nearly expected allele frequencies. We show how intensive selection can be applied using artificial insemination to rapidly generate a multi-locus introgression line and provide recommendations for future breeding of introgression lines. This confirmed introgression line will facilitate later detailed studies of the effects of genetic interactions on complex traits in this population, including growth, and body-composition traits.

Estimation and interpretation of genetic effects with epistasis using the NOIA model

We introduce this communication with a brief outline of the historical landmarks in genetic modeling, especially concerning epistasis. Then, we present methods for the use of genetic modeling in QTL analyses. In particular, we summarize the essential expressions of the natural and orthogonal interactions (NOIA) model of genetic effects. Our motivation for reviewing that theory here is twofold. First, this review presents a digest of the expressions for the application of the NOIA model, which are often mixed with intermediate and additional formulae in the original articles. Second, we make the required theory handy for the reader to relate the genetic concepts to the particular mathematical expressions underlying them. We illustrate those relations by providing graphical interpretations and a diagram summarizing the key features for applying genetic modeling with epistasis in comprehensive QTL analyses. Finally, we briefly review some examples of the application of NOIA to real data and the way it improves the interpretability of the results.

Hierarchical likelihood opens a new way of estimating genetic values using genome-wide dense marker maps

Background

Genome-wide dense markers have been used to detect genes and estimate relative genetic values. Among many methods, Bayesian techniques have been widely used and shown to be powerful in genome-wide breeding value estimation and association studies. However, computation is known to be intensive under the Bayesian framework, and specifying a prior distribution for each parameter is always required for Bayesian computation. We propose the use of hierarchical likelihood to solve such problems.

Results

Using double hierarchical generalized linear models, we analyzed the simulated dataset provided by the QTLMAS 2010 workshop. Marker-specific variances estimated by double hierarchical generalized linear models identified the QTL with large effects for both the quantitative and binary traits. The QTL positions were detected with very high accuracy. For young individuals without phenotypic records, the true and estimated breeding values had Pearson correlation of 0.60 for the quantitative trait and 0.72 for the binary trait, where the quantitative trait had a more complicated genetic architecture involving imprinting and epistatic QTL.

Conclusions

Hierarchical likelihood enables estimation of marker-specific variances under the likelihoodist framework. Double hierarchical generalized linear models are powerful in localizing major QTL and computationally fast.

qtl.outbred: Interfacing outbred line cross data with the R/qtl mapping software

Background

qtl.outbred is an extendible interface in the statistical environment, R, for combining quantitative trait loci (QTL) mapping tools. It is built as an umbrella package that enables outbred genotype probabilities to be calculated and/or imported into the software package R/qtl.

Findings

Using qtl.outbred, the genotype probabilities from outbred line cross data can be calculated by interfacing with a new and efficient algorithm developed for analyzing arbitrarily large datasets (included in the package) or imported from other sources such as the web-based tool, GridQTL.

Conclusion

qtl.outbred will improve the speed for calculating probabilities and the ability to analyse large future datasets. This package enables the user to analyse outbred line cross data accurately, but with similar effort than inbred line cross data.

Targeted resequencing of a genomic region influencing tameness and aggression reveals multiple signals of positive selection

The identification of the causative genetic variants in quantitative trait loci (QTL) influencing phenotypic traits is challenging, especially in crosses between outbred strains. We have previously identified several QTL influencing tameness and aggression in a cross between two lines of wild-derived, outbred rats (Rattus norvegicus) selected for their behavior towards humans. Here, we use targeted sequence capture and massively parallel sequencing of all genes in the strongest QTL in the founder animals of the cross. We identify many novel sequence variants, several of which are potentially functionally relevant. The QTL contains several regions where either the tame or the aggressive founders contain no sequence variation, and two regions where alternative haplotypes are fixed between the founders. A re-analysis of the QTL signal showed that the causative site is likely to be fixed among the tame founder animals, but that several causative alleles may segregate among the aggressive founder animals. Using a formal test for the detection of positive selection, we find 10 putative positively selected regions, some of which are close to genes known to influence behavior. Together, these results show that the QTL is probably not caused by a single selected site, but may instead represent the joint effects of several sites that were targets of polygenic selection.

Fine mapping and replication of QTL in outbred chicken advanced intercross lines

Background

Linkage mapping is used to identify genomic regions affecting the expression of complex traits. However, when experimental crosses such as F₂ populations or backcrosses are used to map regions containing a Quantitative Trait Locus (QTL), the size of the regions identified remains quite large, i.e. 10 or more Mb. Thus, other experimental strategies are needed to refine the QTL locations. Advanced Intercross Lines (AIL) are produced by repeated intercrossing of F₂ animals and successive generations, which decrease linkage disequilibrium in a controlled manner. Although this approach is seen as promising, both to replicate QTL analyses and fine-map QTL, only a few AIL datasets, all originating from inbred founders, have been reported in the literature.

Methods

We have produced a nine-generation AIL pedigree (n = 1529) from two outbred chicken lines divergently selected for body weight at eight weeks of age. All animals were weighed at eight weeks of age and genotyped for SNP located in nine genomic regions where significant or suggestive QTL had previously been detected in the F₂ population. In parallel, we have developed a novel strategy to analyse the data that uses both genotype and pedigree information of all AIL individuals to replicate the detection of and fine-map QTL affecting juvenile body weight.

Results

Five of the nine QTL detected with the original F₂ population were confirmed and fine-mapped with the AIL, while for the remaining four, only suggestive evidence of their existence was obtained. All original QTL were confirmed as a single locus, except for one, which split into two linked QTL.

Conclusions

Our results indicate that many of the QTL, which are genome-wide significant or suggestive in the analyses of large intercross populations, are true effects that can be replicated and fine-mapped using AIL. Key factors for success are the use of large populations and powerful statistical tools. Moreover, we believe that the statistical methods we have developed to efficiently study outbred AIL populations will increase the number of organisms for which in-depth complex traits can be analyzed.

SNP detection and prediction of variability between chicken lines using genome resequencing of DNA pools

BACKGROUND

Next-generation sequencing technologies are widely used for detection of millions of Single Nucleotide Polymorphisms (SNPs) and also provide a means of assessing their variation. This information is useful for composing subsets of highly informative SNPs for region-specific or genome-wide analysis and to identify mutations regulating phenotypic differences within or between populations. In this study, we investigated the sensitivity of SNP detection and introduced the flanking SNPs value (FSV) as a novel measure for predicting SNP-variability using ~5X genome resequencing with ABI SOLID and DNA pools from two chicken lines divergently selected for juvenile bodyweight.

RESULTS

Genotyping with a 60 K SNP chip revealed polymorphisms within or between two divergently selected chicken lines for 31 363 SNPs, 48% of which were also detected using resequencing of DNA pools. SNP detection using resequencing was more powerful for positions with larger differences in allele frequency between the lines. About 50% of the SNPs with non-reference allele frequencies in the range 0.5-0.6 and 67% of those with frequencies > 0.9 could be detected. On average, ~3.7 SNPs/kb were detected by resequencing, with about 5% lower density on microchromosomes than on macrochromosomes. There was a positive correlation between the observed between-line SNP variation from the 60 K chip analysis and our proposed FSV score computed from the genome resequencing data. The strongest correlations on macrochromosomes and microchromosomes were observed when the FSV was calculated with total flanking regions of 62 kb (correlation 0.55) and 38 kb (correlation 0.45), respectively.

CONCLUSIONS

Genome resequencing with limited coverage (~5X) using pooled DNA samples and three non-reference reads as a threshold for SNP detection, identified 50 - 67% of the 60 K SNPs with a non-reference allele frequency larger than 0.5. The SNP density was around 5% lower on the microchromosomes, most likely because of their higher gene content. Our proposed method to estimate the SNP variation (FSV) uses additional sequence information to better predict SNP informativity. The FSV scores showed higher correlations for SNPs with a larger difference in allele frequency between the populations. The correlation was strongest on macrochromosomes, probably due to a lower recombination rate.

Genetic analysis of metabolic traits in an intercross between body weight-selected chicken lines

A network of four interacting loci has been reported previously to influence growth in two lines of chickens divergently selected for body weight at 56 days of age. Located on chromosomes 3 (Growth4), 4 (Growth6), 7 (Growth9), and 20 (Growth12), they explained nearly half of the difference in body weight at selection age between the two lines. The original study reported effects on body weight and fat deposition, but no attempts were made to explore the effects of the network on other phenotypes measured in the F(2) population. In this study we conducted further analyses to evaluate the specific effects of the four-locus network on other metabolic traits as well as refining results from the original study by including a larger number of genetic markers in the quantitative trait locus (QTL) regions. We confirm the previously described effect of the epistatic network on body weight and show that the network increases the total amount of muscle and fat as well as the weight of the internal organs. The network as a whole did not change the relative content of any studied organs or tissues in the body. There was, however, a significant interaction between the loci on chromosomes 3 and 7 that changed the relative proportion of abdominal fat and breast muscle in the chicken by increasing abdominal fat weight without a corresponding increase in muscle mass.

Genomic analysis of variation in hindlimb musculature of mice from the C57BL/6J and DBA/2J lineage

The precise locations of attachment points of muscle to bone-the origin and insertion sites-are crucial anatomical and functional characteristics that influence locomotor performance. Mechanisms that control the development of these interactions between muscle, tendon, and bone are currently not well understood. In a subset of BXD recombinant inbred (RI) strains derived from the C57BL/6J and DBA/2J strains, we observed a soleus femoral attachment anomaly (SFAA) that was rare in both parental strains (Lionikas, Glover et al. 2006). The aim of the present study was to assess suitability of SFAA as a model to study the genetic mechanisms underlying variation in musculoskeletal anatomy. We scored the incidence of SFAA in 55 BXD strains (n = 9 to 136, median = 26, phenotyped animals per strain, for a total number of 2367). Seven strains (BXD1, 12, 38, 43, 48, 54, and 56) exhibited a high incidence of unilateral SFAA (47-89%), whereas 23 strains scored 0%. Exploration of the mechanisms underlying SFAA in 2 high incidence strains, BXD1 and BXD38, indicated that SFAA-relevant genes are to be found in both C57BL/6J and DBA/2J regions of the BXD1 genome. However, not all alleles relevant for the expression of the phenotype were shared between the 2 high-incidence BXD strains. In conclusion, the anatomical origin of the soleus muscle in mouse is controlled by a polygenic system. A panel of BXD RI strains is a useful tool in exploring the genetic mechanisms underlying SFAA and improving our understanding of musculoskeletal development.

A genetic algorithm based method for stringent haplotyping of family data

Background

The linkage phase, or haplotype, is an extra level of information that in addition to genotype and pedigree can be useful for reconstructing the inheritance pattern of the alleles in a pedigree, and computing for example Identity By Descent probabilities. If a haplotype is provided, the precision of estimated IBD probabilities increases, as long as the haplotype is estimated without errors. It is therefore important to only use haplotypes that are strongly supported by the available data for IBD estimation, to avoid introducing new errors due to erroneous linkage phases.

Results

We propose a genetic algorithm based method for haplotype estimation in family data that includes a stringency parameter. This allows the user to decide the error tolerance level when inferring parental origin of the alleles. This is a novel feature compared to existing methods for haplotype estimation. We show that using a high stringency produces haplotype data with few errors, whereas a low stringency provides haplotype estimates in most situations, but with an increased number of errors.

Conclusion

By including a stringency criterion in our haplotyping method, the user is able to maintain the error rate at a suitable level for the particular study; one can select anything from haplotyped data with very small proportion of errors and a higher proportion of non-inferred haplotypes, to data with phase estimates for every marker, when haplotype errors are tolerable. Giving this choice makes the method more flexible and useful in a wide range of applications as it is able to fulfil different requirements regarding the tolerance for haplotype errors, or uncertain marker-phases.

Genetic analysis of an F(2) intercross between two chicken lines divergently selected for body-weight

BACKGROUND

We have performed Quantitative Trait Loci (QTL) analysis of an F(2) intercross between two chicken lines divergently selected for juvenile body-weight. In a previous study 13 identified loci with effects on body-weight, only explained a small proportion of the large variation in the F(2) population. Epistatic interaction analysis however, indicated that a network of interacting loci with large effect contributed to the difference in body-weight of the parental lines. This previous analysis was, however, based on a sparse microsatellite linkage map and the limited coverage could have affected the main conclusions. Here we present a revised QTL analysis based on a high-density linkage map that provided a more complete coverage of the chicken genome. Furthermore, we utilized genotype data from ~13,000 SNPs to search the genome for potential selective sweeps that have occurred in the selected lines.

RESULTS

We constructed a linkage map comprising 434 genetic markers, covering 31 chromosomes but leaving seven microchromosomes uncovered. The analysis showed that seven regions harbor QTL that influence growth. The pair-wise interaction analysis identified 15 unique QTL pairs and notable is that nine of those involved interactions with a locus on chromosome 7, forming a network of interacting loci. The analysis of ~13,000 SNPs showed that a substantial proportion of the genetic variation present in the founder population has been lost in either of the two selected lines since ~60% of the SNPs polymorphic among lines showed fixation in one of the lines. With the current marker coverage and QTL map resolution we did not observe clear signs of selective sweeps within QTL intervals.

CONCLUSION

The results from the QTL analysis using the new improved linkage map are to a large extent in concordance with our previous analysis of this pedigree. The difference in body-weight between the parental chicken lines is caused by many QTL each with a small individual effect. Although the increased chromosomal marker coverage did not lead to the identification of additional QTL, we were able to refine the localization of QTL. The importance of epistatic interaction as a mechanism contributing significantly to the remarkable selection response was further strengthened because additional pairs of interacting loci were detected with the improved map.

Comparison of analyses of the QTLMAS XII common dataset. I: Genomic selection

A dataset was simulated and distributed to participants of the QTLMAS XII workshop who were invited to develop genomic selection models. Each contributing group was asked to describe the model development and validation as well as to submit genomic predictions for three generations of individuals, for which they only knew the genotypes. The organisers used these genomic predictions to perform the final validation by comparison to the true breeding values, which were known only to the organisers. Methods used by the 5 groups fell in 3 classes 1) fixed effects models 2) BLUP models, and 3) Bayesian MCMC based models. The Bayesian analyses gave the highest accuracies, followed by the BLUP models, while the fixed effects models generally had low accuracies and large error variance. The best BLUP models as well as the best Bayesian models gave unbiased predictions. The BLUP models are clearly sensitive to the assumed SNP variance, because they do not estimate SNP variance, but take the specified variance as the true variance. The current comparison suggests that Bayesian analyses on haplotypes or SNPs are the most promising approach for Genomic selection although the BLUP models may provide a computationally attractive alternative with little loss of efficiency. On the other hand fixed effect type models are unlikely to provide any gain over traditional pedigree indexes for selection.

Comparison of analyses of the QTLMAS XII common dataset. II: genome-wide association and fine mapping

As part of the QTLMAS XII workshop, a simulated dataset was distributed and participants were invited to submit analyses of the data based on genome-wide association, fine mapping and genomic selection. We have evaluated the findings from the groups that reported fine mapping and genome-wide association (GWA) efforts to map quantitative trait loci (QTL). Generally the power to detect QTL was high and the Type 1 error was low. Estimates of QTL locations were generally very accurate. Some methods were much better than others at estimating QTL effects, and with some the accuracy depended on simulated effect size or minor allele frequency. There were also indications of bias in the effect estimates. No epistasis was simulated, but the two studies that included searches for epistasis reported several interacting loci, indicating a problem with controlling the Type I error rate in these analyses. Although this study is based on a single dataset, it indicates that there is a need to improve fine mapping and GWA methods with respect to estimation of genetic effects, appropriate choice of significance thresholds and analysis of epistasis.

Evolutionary potential of hidden genetic variation

The ability of a population to respond to natural or artificial selection pressures is determined by the genetic architecture of the selected trait. It is now widely acknowledged that a substantial part of genetic variability can be buffered or released as the result of complex genetic interactions. However, the impact of hidden genetic diversity on phenotypic evolution is still not clear. Here, we argue that a common term to describe the impact of hidden genetic variation on phenotypic change is needed and will help to provide new insights into the contribution of different components of genetic architectures to the evolvability of a character. We introduce the 'genetic charge' concept, to describe how the architecture of a trait can be 'charged' with potential for evolutionary change that can later be 'discharged' in response to selection.

Defining the assumptions underlying modeling of epistatic QTL using variance component methods

Variance component models are commonly used to detect quantitative trait loci (QTL) in general pedigrees. The variance-covariance structure of the random QTL effect is given by the identity by descent (IBD) between genotypes. Epistatic effects have previously been modeled, both for unlinked and linked loci, as a random effect with a variance-covariance structure given by the Hadamard product between the IBD matrices of the direct QTL effects. In the original papers, the model was given but not derived. Here, we identify the underlying assumptions of this previously proposed model. It assumes that either an unlinked QTL or a fully informative marker (i.e., all marker alleles are unique in the base generation) is located between the loci. We discuss the need of developing a general algorithm to estimate the variance-covariance structure of the random epistatic effect for linked loci.

Phenotypic evolution from genetic polymorphisms in a radial network architecture

Background

The genetic architecture of a quantitative trait influences the phenotypic response to natural or artificial selection. One of the main objectives of genetic mapping studies is to identify the genetic factors underlying complex traits and understand how they contribute to phenotypic expression. Presently, we are good at identifying and locating individual loci with large effects, but there is a void in describing more complex genetic architectures. Although large networks of connected genes have been reported, there is an almost complete lack of information on how polymorphisms in these networks contribute to phenotypic variation and change. To date, most of our understanding comes from theoretical, model-based studies, and it remains difficult to assess how realistic their conclusions are as they lack empirical support.

Results

A previous study provided evidence that nearly half of the difference in eight-week body weight between two divergently selected lines of chickens was a result of four loci organized in a 'radial' network (one central locus interacting with three 'radial' loci that, in turn, only interacted with the central locus). Here, we study the relationship between phenotypic change and genetic polymorphism in this empirically detected network. We use a model-free approach to study, through individual-based simulations, the dynamic properties of this polymorphic and epistatic genetic architecture. The study provides new insights to how epistasis can modify the selection response, buffer and reveal effects of major loci leading to a progressive release of genetic variation. We also illustrate the difficulty of predicting genetic architecture from observed selection response, and discuss mechanisms that might lead to misleading conclusions on underlying genetic architectures from quantitative trait locus (QTL) experiments in selected populations.

Conclusion

Considering both molecular (QTL) and phenotypic (selection response) data, as suggested in this work, provides additional insights into the genetic mechanisms involved in the response to selection. Such dissection of genetic architectures and in-depth studies of their ability to contribute to short- or long-term selection response represents an important step towards a better understanding of the genetic bases of complex traits and, consequently, of the evolutionary properties of populations.

A general and efficient method for estimating continuous IBD functions for use in genome scans for QTL

BACKGROUND

Identity by descent (IBD) matrix estimation is a central component in mapping of Quantitative Trait Loci (QTL) using variance component models. A large number of algorithms have been developed for estimation of IBD between individuals in populations at discrete locations in the genome for use in genome scans to detect QTL affecting various traits of interest in experimental animal, human and agricultural pedigrees. Here, we propose a new approach to estimate IBD as continuous functions rather than as discrete values.

RESULTS

Estimation of IBD functions improved the computational efficiency and memory usage in genome scanning for QTL. We have explored two approaches to obtain continuous marker-bracket IBD-functions. By re-implementing an existing and fast deterministic IBD-estimation method, we show that this approach results in IBD functions that produces the exact same IBD as the original algorithm, but with a greater than 2-fold improvement of the computational efficiency and a considerably lower memory requirement for storing the resulting genome-wide IBD. By developing a general IBD function approximation algorithm, we show that it is possible to estimate marker-bracket IBD functions from IBD matrices estimated at marker locations by any existing IBD estimation algorithm. The general algorithm provides approximations that lead to QTL variance component estimates that even in worst-case scenarios are very similar to the true values. The approach of storing IBD as polynomial IBD-function was also shown to reduce the amount of memory required in genome scans for QTL.

CONCLUSION

In addition to direct improvements in computational and memory efficiency, estimation of IBD-functions is a fundamental step needed to develop and implement new efficient optimization algorithms for high precision localization of QTL. Here, we discuss and test two approaches for estimating IBD functions based on existing IBD estimation algorithms. Our approaches provide immediately useful techniques for use in single QTL analyses in the variance component QTL mapping framework. They will, however, be particularly useful in genome scans for multiple interacting QTL, where the improvements in both computational and memory efficiency are the key for successful development of efficient optimization algorithms to allow widespread use of this methodology.

A unified model for functional and statistical epistasis and its application in quantitative trait Loci analysis

Interaction between genes, or epistasis, is found to be common and it is a key concept for understanding adaptation and evolution of natural populations, response to selection in breeding programs, and determination of complex disease. Currently, two independent classes of models are used to study epistasis. Statistical models focus on maintaining desired statistical properties for detection and estimation of genetic effects and for the decomposition of genetic variance using average effects of allele substitutions in populations as parameters. Functional models focus on the evolutionary consequences of the attributes of the genotype-phenotype map using natural effects of allele substitutions as parameters. Here we provide a new, general and unified model framework: the natural and orthogonal interactions (NOIA) model. NOIA implements tools for transforming genetic effects measured in one population to the ones of other populations (e.g., between two experimental designs for QTL) and parameters of statistical and functional epistasis into each other (thus enabling us to obtain functional estimates of QTL), as demonstrated numerically. We develop graphical interpretations of functional and statistical models as regressions of the genotypic values on the gene content, which illustrates the difference between the models--the constraint on the slope of the functional regression--and when the models are equivalent. Furthermore, we use our theoretical foundations to conceptually clarify functional and statistical epistasis, discuss the advantages of NOIA over previous theory, and stress the importance of linking functional and statistical models.

Increasing the efficiency of variance component quantitative trait loci analysis by using reduced-rank identity-by-descent matrices

Recent technological development in genetics has made large-scale marker genotyping fast and practicable, facilitating studies for detection of QTL in large general pedigrees. We developed a method that speeds up restricted maximum-likelihood (REML) algorithms for QTL analysis by simplifying the inversion of the variance-covariance matrix of the trait vector. The method was tested in an experimental chicken pedigree including 767 phenotyped individuals and 14 genotyped markers on chicken chromosome 1. The computation time in a chromosome scan covering 475 cM was reduced by 43% when the analysis was based on linkage only and by 72% when linkage disequilibrium information was included. The relative advantage of using our method increases with pedigree size, marker density, and linkage disequilibrium, indicating even greater improvements in the future.

Statistical epistasis is a generic feature of gene regulatory networks

Functional dependencies between genes are a defining characteristic of gene networks underlying quantitative traits. However, recent studies show that the proportion of the genetic variation that can be attributed to statistical epistasis varies from almost zero to very high. It is thus of fundamental as well as instrumental importance to better understand whether different functional dependency patterns among polymorphic genes give rise to distinct statistical interaction patterns or not. Here we address this issue by combining a quantitative genetic model approach with genotype-phenotype models capable of translating allelic variation and regulatory principles into phenotypic variation at the level of gene expression. We show that gene regulatory networks with and without feedback motifs can exhibit a wide range of possible statistical genetic architectures with regard to both type of effect explaining phenotypic variance and number of apparent loci underlying the observed phenotypic effect. Although all motifs are capable of harboring significant interactions, positive feedback gives rise to higher amounts and more types of statistical epistasis. The results also suggest that the inclusion of statistical interaction terms in genetic models will increase the chance to detect additional QTL as well as functional dependencies between genetic loci over a broad range of regulatory regimes. This article illustrates how statistical genetic methods can fruitfully be combined with nonlinear systems dynamics to elucidate biological issues beyond reach of each methodology in isolation.

Epistatic regulation of behavioural and morphological traits in the zebrafish (Danio rerio)

There is currently a lack of studies examining epistasis in general, and specifically for behavioural traits of evolutionary significance. The advent of more efficient analytical methods for exploring epistasis in QTL studies removes the computational restraint on this type of analysis and suggests that performing further analyses of existing datasets may reveal a more complete picture of the genetic architecture of the traits. Here we report the results from an epistatic QTL analysis of an F2 cross between a wild population and a standard laboratory strain of zebrafish. This further analysis was performed using a simultaneous search to identify epistatically interacting QTL affecting behavioural and morphological traits and uncovered several novel epistatic interactions that reached either genome-wide or suggestive significance levels as determined by a randomisation testing approach. These results provide novel insight into the genetic architecture of the regulation of behavioural as well as morphological phenotypes and call for more studies of epistasis for this group of traits.

Epistasis and the release of genetic variation during long-term selection

It is an enigma how long-term selection in model organisms and agricultural species can lead to marked phenotypic changes without exhausting genetic variation for the selected trait. Here, we show that the genetic architecture of an apparently major locus for growth in chicken dissects into a genetic network of four interacting loci. The interactions in this radial network mediate a considerably larger selection response than predicted by a single-locus model.

Methodological aspects of the genetic dissection of gene expression

MOTIVATION

Dissection of the genetics underlying gene expression utilizes techniques from microarray analyses as well as quantitative trait loci (QTL) mapping. Available QLT mapping methods are not tailored for the highly automated analyses required to deal with the thousand of gene transcripts encountered in the mapping of QTL affecting gene expression (sometimes referred to as eQTL). This report focuses on the adaptation of QTL mapping methodology to perform automated mapping of QTL affecting gene expression.

RESULTS

The analyses of expression data on > 12,000 gene transcripts in BXD recombinant inbred mice found, on average, 629 QTL exceeding the genome-wide 5% threshold. Using additional information on trait repeatabilities and QTL location, 168 of these were classified as 'high confidence' QTL. Current sample sizes of genetical genomics studies make it possible to detect a reasonable number of QTL using simple genetic models, but considerably larger studies are needed to evaluate more complex genetic models. After extensive analyses of real data and additional simulated data (altogether > 300,000 genome scans) we make the following recommendations for detection of QTL for gene expression: (1) For populations with an unbalanced number of replicates on each genotype, weighted least squares should be preferred above ordinary least squares. Weights can be based on repeatability of the trait and the number of replicates. (2) A genome scan based on multiple marker information but analysing only at marker locations is a good approximation to a full interval mapping procedure. (3) Significance testing should be based on empirical genome-wide significance thresholds that are derived for each trait separately. (4) The significant QTL can be separated into high and low confidence QTL using a false discovery rate that incorporates prior information such as transcript repeatabilities and co-localization of gene-transcripts and QTL. (5) Including observations on the founder lines in the QTL analysis should be avoided as it inflates the test statistic and increases the Type I error. (6) To increase the computational efficiency of the study, use of parallel computing is advised. These recommendations are summarized in a possible strategy for mapping of QTL in a least squares framework.

AVAILABILITY

The software used for this study is available on request from the authors.

Simultaneous mapping of epistatic QTL in DU6i × DBA/2 mice

We have mapped epistatic quantitative trait loci (QTL) in an F2 cross between DU6i x DBA/2 mice. By including these epistatic QTL and their interaction parameters in the genetic model, we were able to increase the genetic variance explained substantially (8.8%-128.3%) for several growth and body composition traits. We used an analysis method based on a simultaneous search for epistatic QTL pairs without assuming that the QTL had any effect individually. We were able to detect several QTL that could not be detected in a search for marginal QTL effects because the epistasis cancelled out the individual effects of the QTL. In total, 23 genomic regions were found to contain QTL affecting one or several of the traits and eight of these QTL did not have significant individual effects. We identified 44 QTL pairs with significant effects on the traits, and, for 28 of the pairs, an epistatic QTL model fit the data significantly better than a model without interactions. The epistatic pairs were classified by the significance of the epistatic parameters in the genetic model, and visual inspection of the two-locus genotype means identified six types of related genotype-phenotype patterns among the pairs. Five of these patterns resembled previously published patterns of QTL interactions.

The genetic dissection of immune response using gene-expression studies and genome mapping

Functional genomics has been applied to the genetic dissection of immune response in different ways: (1) experimental crosses between lines that differ in their (non-) specific immune response have been used to detect quantitative trait loci (QTL) underlying these differences. (2) The measurement of gene expression levels for thousands of genes using microarrays or oligonucleotide chips to identify differential expression with regard to antigen challenge: (a) before and after infection, (b) resistant versus susceptible lines, or (c) combinations of both. Interpretation of QTL results is hampered by the fact that confidence regions of the QTL are large and can contain hundreds of potential candidate genes for the QTL. At the same time, the microarray experiments tend to show large numbers of differentially expressed genes without identifying the relationships between these genes. In the recently proposed 'genetical genomics' framework, members of a segregating population are characterised for genome-wide molecular markers and for gene expression levels. This facilitates the mapping of expression-QTL (eQTL): loci in the genome that control the expression of genes. Initial applications of this approach are critically reviewed and potential applications of this approach with regard to immune response are presented.

Simultaneous mapping of epistatic QTL in chickens reveals clusters of QTL pairs with similar genetic effects on growth

We used simultaneous mapping of interacting quantitative trait locus (QTL) pairs to study various growth traits in a chicken F2 intercross. The method was shown to increase the number of detected QTLs by 30 % compared with a traditional method detecting QTLs by their marginal genetic effects. Epistasis was shown to be an important contributor to the genetic variance of growth, with the largest impact on early growth (before 6 weeks of age). There is also evidence for a discrete set of interacting loci involved in early growth, supporting the previous findings of different genetic regulation of early and late growth in chicken. The genotype-phenotype relationship was evaluated for all interacting QTL pairs and 17 of the 21 evaluated QTL pairs could be assigned to one of four clusters in which the pairs in a cluster have very similar genetic effects on growth. The genetic effects of the pairs indicate commonly occurring dominance-by-dominance, heterosis and multiplicative interactions. The results from this study clearly illustrate the increase in power obtained by using this novel method for simultaneous detection of epistatic QTL, and also how visualization of genotype-phenotype relationships for epistatic QTL pairs provides new insights to biological mechanisms underlying complex traits.

Chicken genomics: feather-pecking and victim pigmentation

Feather-pecking in domestic birds is associated with cannibalism and severe welfare problems. It is a dramatic example of a spiteful behaviour in which the victim's fitness is reduced for no immediate direct benefit to the perpetrator and its evolution is unexplained. Here we show that the plumage pigmentation of a chicken may predispose it to become a victim: birds suffer more drastic feather-pecking when the colour of their plumage is due to the expression of a wild recessive allele at PMEL17, a gene that controls plumage melanization, and when these birds are relatively common in a flock. These findings, obtained using an intercross between a domestic fowl and its wild ancestor, have implications for the welfare of domestic species and offer insight into the genetic changes associated with the evolution of feather-pecking during the early stages of domestication.

Simultaneous search for multiple QTL using the global optimization algorithm DIRECT

MOTIVATION

A simultaneous search is necessary for maximizing the power to detect epistatic quantitative trait loci (QTL). The computational complexity demands that the traditional exhaustive search be replaced by a more efficient global optimization algorithm.

RESULTS

We have the previously known algorithm adapted DIRECT, to the problem of simultaneous mapping of multiple QTL. We have compared DIRECT with standard exhaustive search and a genetic algorithm previously used for QTL mapping in two dimensions. In all two- and three-QTL test cases, DIRECT accurately finds the global optimum two to four orders of magnitude faster than when using an exhaustive search, and one order of magnitude faster than when using the genetic algorithm. Thus, randomization testing for determining empirical significance thresholds for at least three QTL is made feasible by the use of DIRECT.

AVAILABILITY

The code of the prototype implementation is available at http://user.it.uu.se/~kl/qtl_software.html

Efficient algorithms for quantitative trait loci mapping problems

Rapid advances in molecular genetics push the need for efficient data analysis. Advanced algorithms are necessary for extracting all possible information from large experimental data sets. We present a general linear algebra framework for quantitative trait loci (QTL) mapping, using both linear regression and maximum likelihood estimation. The formulation simplifies future comparisons between and theoretical analyses of the methods. We show how the common structure of QTL analysis models can be used to improve the kernel algorithms, drastically reducing the computational effort while retaining the original analysis results. We have evaluated our new algorithms on data sets originating from two large F(2) populations of domestic animals. Using an updating approach, we show that 1-3 orders of magnitude reduction in computational demand can be achieved for matrix factorizations. For interval-mapping/composite-interval-mapping settings using a maximum likelihood model, we also show how to use the original EM algorithm instead of the ECM approximation, significantly improving the convergence and further reducing the computational time. The algorithmic improvements makes it feasible to perform analyses which have previously been deemed impractical or even impossible. For example, using the new algorithms, it is reasonable to perform permutation testing using exhaustive search on populations of 200 individuals using an epistatic two-QTL model.

Major Growth QTLs in Fowl Are Related to Fearful Behavior: Possible Genetic Links Between Fear Responses and Production Traits in a Red Junglefowl × White Leghorn Intercross

The aim of this work was to study fear responses and their relation to production traits in red junglefowl ( Gallus gallus spp.), White Leghorn ( Gallus domesticus ), and their F2-progeny. Quantitative trait locus (QTL) analyses were performed for behavioral traits to gain information about possible genetic links between fear-related behaviors and production. Four behavioral tests were performed that induce different levels of acute fear (open field [OF], exposure to a novel object, tonic immobility, and restraint). Production traits, that is, egg production, sexual maturity (in females), food intake, and growth, were measured individually. A genome scan using 105 microsatellite markers was carried out to identify QTLs controlling the traits studied. In the OF and novel object tests (NO), Leghorns showed less fear behavior than junglefowl, whereas junglefowl behaved less fearfully in the tonic immobility test (TI) and were more active in the restraint test. In the F2 progeny, only weak phenotypic associations were found between production traits and fear behavior. A significant QTL for TI duration was found on chromosome 1 that coincided with a QTL for egg weight and growth in the same animals. Another QTL for NO in males coincided with another major growth QTL. These two known growth QTLs affected a wide range of reactions in different tests. Several other significant and suggestive QTLs for behavioral traits related to fear were found. These QTLs did not coincide with QTLs for production traits, indicating that these fear variables may not be genetically linked to the production traits we measured here. The results show that loci affecting important production traits are located in the same chromosomal region as loci affecting different fear-related behaviors.

The twofold difference in adult size between the red junglefowl and White Leghorn chickens is largely explained by a limited number of QTLs

A large intercross between the domestic White Leghorn chicken and the wild ancestor, the red junglefowl, has been used in a Quantitative Trait Loci (QTL) study of growth and egg production. The linkage map based on 105 marker loci was in good agreement with the chicken consensus map. The growth of the 851 F2 individuals was lower than both parental lines prior to 46 days of age and intermediate to the two parental lines thereafter. The QTL analysis of growth traits revealed 13 loci that showed genome-wide significance. The four major growth QTLs explained 50 and 80% of the difference in adult body weight between the founder populations for females and males, respectively. A major QTL for growth, located on chromosome 1 appears to have pleiotropic effects on feed consumption, egg production and behaviour. There was a strong positive correlation between adult body weight and average egg weight. However, three QTLs affecting average egg weight but not body weight were identified. An interesting observation was that the estimated effects for the four major growth QTLs all indicated a codominant inheritance.

A global search reveals epistatic interaction between QTL for early growth in the chicken

We have identified quantitative trait loci (QTL) explaining a large proportion of the variation in body weights at different ages and growth between chronological ages in an F(2) intercross between red junglefowl and White Leghorn chickens. QTL were mapped using forward selection for loci with significant marginal genetic effects and with a simultaneous search for epistatic QTL pairs. We found 22 significant loci contributing to these traits, nine of these were only found by the simultaneous two-dimensional search, which demonstrates the power of this approach for detecting loci affecting complex traits. We have also estimated the relative contribution of additive, dominance, and epistasis effects to growth and the contribution of epistasis was more pronounced prior to 46 days of age, whereas additive genetic effects explained the major portion of the genetic variance later in life. Several of the detected loci affected either early or late growth but not both. Very few loci affected the entire growth process, which points out that early and late growth, at least to some extent, have different genetic regulation.

QTL analysis of a red junglefowl x White Leghorn intercross reveals trade-off in resource allocation between behavior and production traits

Behaviors with high energetic costs may decrease in frequency in domestic animals as a response to selection for increased production. The aim of this study was to quantify production traits, foraging behavior, and social motivation in F2 progeny from a White Leghorn x red junglefowl intercross (n = 751-1046) and to perform QTL analyses on the behavioral traits. A foraging-social maze was used for behavioral testing, which consisted of four identical arms and a central box. In two arms there was ad libitum access to the birds' usual food, and in the other two there was novel food (sunflower seeds) mixed with cat litter. In one arm with each of the two food sources, social stimuli were simulated by the presence of a mirror. Each bird could therefore feed on novel or well known food either alone or in the perceived company of a conspecific. Egg production, sexual maturity (females), food intake, and growth were measured individually, and residual food intake and metabolic body weight were estimated using standard methods. A genome scan using 104 microsatellite markers was carried out to identify QTLs affecting behavioral traits. Phenotypic growth rates at different ages showed weak associations in both sexes. Sexual maturity and egg weight were not strongly correlated to growth, indicating that these traits are not genetically linked. Time spent in each arm and in the central part of the maze was analyzed using principal component analyses. Four principal components (PC) were extracted, each reflecting a pattern of behavior in the maze. Females with early onset of sexual maturity scored higher on the PC1 reflecting preference for free food without social stimuli, and females with higher egg production scored higher on the PC2 reflecting exploration. Males with an overall higher growth rate and higher residual food intake scored higher on the PC3, which possibly reflected fear of the test situation, and tended to score higher on the PC4 reflecting low contrafreeloading. Significant QTLs were found for PC1 and PC4 scores on chromosomes 27 and 7, respectively. The location of the QTLs coincided with known QTLs for growth rate and body weight. The results suggest a trade-off between energy-demanding behavior and high production and that some of this may be caused by genetic linkage or pleiotropic gene effects.

Use of randomization testing to detect multiple epistatic QTLs

Here, we describe a randomization testing strategy for mapping interacting quantitative trait loci (QTLs). In a forward selection strategy, non-interacting QTLs and simultaneously mapped interacting QTL pairs are added to a total genetic model. Simultaneous mapping of epistatic QTLs increases the power of the mapping strategy by allowing detection of interacting QTL pairs where none of the QTL can be detected by their marginal additive and dominance effects. Randomization testing is used to derive empirical significance thresholds for every model selection step in the procedure. A simulation study was used to evaluate the statistical properties of the proposed randomization tests and for which types of epistasis simultaneous mapping of epistatic QTLs adds power. Least squares regression was used for QTL parameter estimation but any other QTL mapping method can be used. A genetic algorithm was used to search for interacting QTL pairs, which makes the proposed strategy feasible for single processor computers. We believe that this method will facilitate the evaluation of the importance at epistatic interaction among QTLs controlling multifactorial traits and disorders.

Melanocortin-4 receptor (MC4R) genotypes have no major effect on fatness in a Large White x Wild Boar intercross

The melanocortin-4 receptor (MC4R), a G-protein coupled receptor, is implicated in mediating the effect of leptin on food intake and energy balance. A previous candidate gene study reported an association between an MC4R missense mutation (Asp298Asn) and fatness, growth and feed intake in pigs. To assess this association further, we analysed the segregation of this missense mutation in relation to variation in fatness traits using a Wild Boar x Large White intercross. The Wild Boar and Large White founders were homozygous for different MC4R alleles. The MC4R was assigned to the expected region on pig chromosome 1. The statistical evaluation did not reveal any indication of a significant effect on fatness related traits in this pedigree.

Parallel computing in interval mapping of quantitative trait loci

Linear regression analysis is considered the least computationally demanding method for mapping quantitative trait loci (QTL). However, simultaneous search for multiple QTL, the use of permutations to obtain empirical significance thresholds, and larger experimental studies significantly increase the computational demand. This report describes an easily implemented parallel algorithm, which significantly reduces the computing time in both QTL mapping and permutation testing. In the example provided, the analysis time was decreased to less than 15% of a single processor system by the use of 18 processors. We indicate how the efficiency of the analysis could be improved by distributing the computations more evenly to the processors and how other ways of distributing the data facilitate the use of more processors. The use of parallel computing in QTL mapping makes it possible to routinely use permutations to obtain empirical significance thresholds for multiple traits and multiple QTL models. It could also be of use to improve the computational efficiency of the more computationally demanding QTL analysis methods.

The use of a genetic algorithm for simultaneous mapping of multiple interacting quantitative trait loci

Here we describe a general method for improving computational efficiency in simultaneous mapping of multiple interacting quantitative trait loci (QTL). The method uses a genetic algorithm to search for QTL in the genome instead of an exhaustive enumerative ("step-by-step") search. It can be used together with any method of QTL mapping based on a genomic search, since it only provides a more efficient way to search the genome for QTL. The computational demand decreases by a factor of approximately 130 when using genetic algorithm-based mapping instead of an exhaustive enumerative search for two QTL in a genome size of 2000 cM using a resolution of 1 cM. The advantage of using a genetic algorithm increases further for larger genomes, higher resolutions, and searches for more QTL. We show that a genetic algorithm-based search has efficiency higher than or equal to a search method conditioned on previously identified QTL for all epistatic models tested and that this efficiency is comparable to that of an exhaustive search for multiple QTL. The genetic algorithm is thus a powerful and computationally tractable alternative to the exhaustive enumerative search for simultaneous mapping of multiple interacting QTL. The use of genetic algorithms for simultaneous mapping of more than two QTL and for determining empirical significance thresholds using permutation tests is also discussed.

The origin of the domestic pig: independent domestication and subsequent introgression

The domestic pig originates from the Eurasian wild boar (Sus scrofa). We have sequenced mitochondrial DNA and nuclear genes from wild and domestic pigs from Asia and Europe. Clear evidence was obtained for domestication to have occurred independently from wild boar subspecies in Europe and Asia. The time since divergence of the ancestral forms was estimated at approximately 500,000 years, well before domestication approximately 9,000 years ago. Historical records indicate that Asian pigs were introduced into Europe during the 18th and early 19th centuries. We found molecular evidence for this introgression and the data indicated a hybrid origin of some major "European" pig breeds. The study is an advance in pig genetics and has important implications for the maintenance and utilization of genetic diversity in this livestock species.

High in situ rat intestinal permeability of artemisinin unaffected by multiple dosing and with no evidence of P-glycoprotein involvement

The objective of this study was to investigate whether the decrease in artemisinin bioavailability after repeated oral dosing in humans can be a result of increased efflux of artemisinin by P-glycoprotein or decreased membrane transport at the intestinal barrier. The effective jejunal permeability (Peff) of artemisinin was investigated using an in situ rat perfusion model. Fifty-four rats were randomized to one of three treatment arms: no pretreatment, pretreatment with artemisinin emulsion for 5 days (60 mg/kg/day, p.o. ), or pretreatment with emulsion vehicle for 5 days. The rats within each treatment arm were randomized further to be jejunally perfused with either low (500 ng/ml) or high (5000 ng/ml) artemisinin concentration or low artemisinin concentration plus the P-glycoprotein inhibitor R,S-verapamil (400 microg/ml). Perfusate samples were assayed for content of artemisinin, R,S-verapamil, and perfusion viability markers. Artemisinin Peff was 1.44 +/- 0.38, 1. 17 +/- 0.32, and 1.71 +/- 0.29 (.10(-4), cm/s) in rats receiving no pretreatment and perfused with low, high, or low artemisinin concentration plus verapamil, respectively. Multiple oral dosing of artemisinin did not affect the jejunal permeability of artemisinin. R,S-verapamil Peff was similar in artemisinin-pretreated rats (1.09 +/- 0.54. 10(-4), cm/s) and rats pretreated with only vehicle (1.07 +/- 0.37. 10(-4), cm/s). The decrease in artemisinin bioavailability after multiple oral dosing in human is probably not a result of changes in P-glycoprotein expression or general intestinal transport. It seems more likely attributed to increased hepatocellular activity. Furthermore, artemisinin exhibits high jejunal permeability and is neither a substrate nor inducer of P-glycoprotein.