Large-scale in silico mapping of complex quantitative traits in inbred mice

Understanding the genetic basis of common disease and disease-related quantitative traits will aid in the development of diagnostics and therapeutics. The processs of gene discovery can be sped up by rapid and effective integration of well-defined mouse genome and phenome data resources. We describe here an in silico gene-discovery strategy through genome-wide association (GWA) scans in inbred mice with a wide range of genetic variation. We identified 937 quantitative trait loci (QTLs) from a survey of 173 mouse phenotypes, which include models of human disease (atherosclerosis, cardiovascular disease, cancer and obesity) as well as behavioral, hematological, immunological, metabolic, and neurological traits. 67% of QTLs were refined into genomic regions <0.5 Mb with approximately 40-fold increase in mapping precision as compared with classical linkage analysis. This makes for more efficient identification of the genes that underlie disease. We have identified two QTL genes, Adam12 and Cdh2, as causal genetic variants for atherogenic diet-induced obesity. Our findings demonstrate that GWA analysis in mice has the potential to resolve multiple tightly linked QTLs and achieve single-gene resolution. These high-resolution QTL data can serve as a primary resource for positional cloning and gene identification in the research community.     

Heritability, correlations and in silico mapping of locomotor behavior and neurochemistry in inbred strains of mice

The midbrain dopamine system mediates normal and pathologic behaviors related to motor activity, attention, motivation/reward and cognition. These are complex, quantitative traits whose variation among individuals is modulated by genetic, epigenetic and environmental factors. Conventional genetic methods have identified several genes important to this system, but the majority of factors contributing to the variation remain unknown. To understand these genetic and environmental factors, we initiated a study measuring 21 behavioral and neurochemical traits in 15 common inbred mouse strains. We report trait data, heritabilities and genetic and non-genetic correlations between pheno-types. In general, the behavioral traits were more heritable than neurochemical traits, and both genetic and non-genetic correlations within these trait sets were high. Surprisingly, there were few significant correlations between the behavioral and the individual neurochemical traits. However, striatal serotonin and one measure of dopamine turnover (DOPAC/DA) were highly correlated with most behavioral measures. The variable accounting for the most variation in behavior was mouse strain and not a specific neurochemical measure, suggesting that additional genetic factors remain to be determined to account for these behavioral differences. We also report the prospective use of the in silico method of quantitative trait loci (QTL) analysis and demonstrate difficulties in the use of this method, which failed to detect significant QTLs for the majority of these traits. These data serve as a framework for further studies of correlations between different midbrain dopamine traits and as a guide for experimental cross designs to identify QTLs and genes that contribute to these traits.     

Multi-species comparative mapping in silico using the COMPASS strategy

MOTIVATION: The completion of human and mouse genome sequences provides a valuable resource for decoding other mammalian genomes. The comparative mapping by annotation and sequence similarity (COMPASS) strategy takes advantage of the resource and has been used in several genome-mapping projects. It uses existing comparative genome maps based on conserved regions to predict map locations of a sequence. An automated multiple-species COMPASS tool can facilitate in the genome sequencing effort and comparative genomics study of other mammalian species. RESULTS: The prerequisite of COMPASS is a comparative map table between the reference genome and the predicting genome. We have built and collected comparative maps among five species including human, cattle, pig, mouse and rat. Cattle-human and pig-human comparative maps were built based on the positions of orthologous markers and the conserved synteny groups between human and cattle and human and pig genomes, respectively. Mouse-human and rat-human comparative maps were based on the conserved sequence segments between the two genomes. With a match to human genome sequences, the approximate location of a query sequence can be predicted in cattle, pig, mouse and rat genomes based on the position of the match relatively to the orthologous markers or the conserved segments. AVAILABILITY: The COMPASS-tool and databases are available at http://titan.biotec.uiuc.edu/COMPASS/     

An AFLP-based genome-wide mapping strategy

To efficiently determine the chromosomal location of phenotypic mutants, we designed a genome-wide mapping strategy that can be used in any crop for which a dense AFLP (Amplified Fragment Length Polymorphism) map is available or can be made. The AFLP technique is particularly suitable to initiate map-based cloning projects because it detects many markers per reaction. First a standard set of AFLP primer combinations that results in a framework of AFLP markers well dispersed over the genome is selected. These primer combinations are applied to a limited number of mutant individuals from a segregating population to register linkage and non-linkage of the AFLP markers to the gene-of-interest. Further delineation of the area of interest is accomplished by analyzing the remaining recombinants and additional mutant individuals with AFLP markers that lie within the identified region. We illustrate the usefulness of the method by mapping three rotunda (ron) leaf-form mutant loci of Arabidopsis thaliana and show that in the initial phase of map-based cloning projects a 400–600 kb interval can be identified for the average mutant locus within a few weeks. Once such an area is identified and before initiating the more time-consuming fine-mapping procedure, it is essential to examine publicly available databases for candidate genes and known mutants in the identified region. The 390-kb interval on chromosome 4 that harbors the ron2 mutation, also carries a known flower mutant, leunig (lug); upon crossing, the two mutants appeared to be allelic. When no such candidates are found, the mapping procedure should be continued. We present a strategy to efficiently select recombinants that can be used for fine mapping.Electronic Supplementary Material Supplementary material is available in the online version of this article at .Communicated by R. Hagemann     

In silico gene selection strategy for custom microarray design

Microarray technology has become an important tool for studying large-scale gene expression for a diversity of biological applications. However, there are a number of experimental settings for which commercial arrays are either unsuitable or unavailable despite the existence of sequence information. With the increasing availability of custom array manufacturing services, it is now feasible to design high-density arrays for any organism having sequence data. However, there have been relatively few reports discussing gene selection, an important first step in array design. Here we propose an in silico strategy for custom microarray gene selection that is applicable to a wide range of organisms, based on utilizing public domain microarray information to interrogate existing sequence data and to identify a set of homologous genes in any organism of interest. We demonstrate the utility of this approach by applying it to the selection of candidate genes for a custom Xenopus laevis microarray. A significant finding of this study is that 3%-4% of Xenopus expressed sequence tags (ESTs) are in an orientation contrary to that indicated in the public database entry (http://mssaha.people.wm.edu/suppMSS.html).     

Gene mapping of sperm quality parameters in recombinant inbred strains of mice

The aim of this study was to map chromosomal regions containing hypothetical genes responsible for the following parameters of mouse semen quality: (1) the percentage of sperm with abnormal head morphology, (2) the level of dead spermatozoa, (3) the percentage of sperm tails with residual cytoplasmic droplets, and (4) the percentage of sperm with impaired sperm tail membrane integrity. We also analyzed any possible correlations between these parameters. The most appropriate animal model for mapping genes controlling quantitative traits (QTL, quantitative trait locus) is a set of recombinant inbred (RI) strains. The set of RI strains used in this study was derived from crosses between two inbred mouse strains, KE and CBA/Kw, which differ significantly in fertility parameters and gamete quality. We analyzed the four parameters of sperm quality in male mice from two parental strains and from 12 RI strains. The strain distribution pattern (SDP) of 187 polymorphic microsatellite markers was prepared for 20 chromosomes of the mouse genome in 12 RI strains. We correlated the SDP of these markers with the values of sperm quality parameters, using MapManager QTX software (ver. b18). The mapping procedure indicated that the percentage of sperm with abnormal head morphology is controlled by gene(s) located in chromosomal regions 11q24, 11q31 and 6q15.6. There was also a strong correlation between male body weight and the hypothetical gene(s) in chromosomal region 18q47. A detailed analysis of the genes located in these regions enabled us to prepare a list of candidate genes. We discuss the basis of the correlation between the measured parameters.     

Power estimation for in silico mapping

A fundamentally new approach to gene mapping of complex traits was suggested recently. It consists in computer analysis of existing databases on the phenotypes and single nucleotide polymorphisms (SNPs) in inbred mouse strains and was termed in silico mapping. The power of this method has been studied by simulating quantitative traits controlled by one, two, or three genes. The results have demonstrated that the power of in silico mapping is high in the case of a monogenic trait. The probability of mapping all genes determining a digenic or, especially, trigenic trait is low. If two or three genes make equal phenotypic contributions to a trait, the proportions of experiments where none of them is localized are 17 and 25%, respectively. In the case of a major gene effect, when the phenotypic contribution of one gene considerably exceeds those of the other genes, the probability to map the major gene is 0.95 and 0.80 for the digenic and trigenic models, respectively. This shows that, in the case of polygenic control, the new method could localize only the genes with major effects, while most genes involved in the control of the trait would not be mapped.     

Power estimation for in silico mapping

A fundamentally new approach to gene mapping of complex traits was suggested recently. It consists in computer analysis of existing databases on the phenotypes and single nucleotide polymorphisms (SNPs) in inbred mouse strains and was termed in silico mapping. The power of this method has been studied by simulating quantitative traits controlled by one, two, or three genes. The results have demonstrated that the power of in silico mapping is high in the case of a monogenic trait. The probability of mapping all genes determining a digenic or, especially, trigenic trait is low. If two or three genes make equal phenotypic contributions to a trait, the proportions of experiments where none of them is localized are 17 and 25%, respectively. In the case of a major gene effect, when the phenotypic contribution of one gene considerably exceeds those of the other genes, the probability to map the major gene is 0.95 and 0.80 for the digenic and trigenic models, respectively. This shows that, in the case of polygenic control, the new method could localize only the genes with major effects, while most genes involved in the control of the trait would not be mapped.     

An optimized predictive strategy for interactome mapping

We present an optimized experimental strategy that can accelerate progress toward identifying the majority of pairwise protein interactions. Our method involves applying a predictive algorithm, based on the existing data, to identify protein pairs likely to interact and prioritizing these for screening. The approach is iterative as additional data allows one to refine predictions directing the next stage of experimentation.     

Genomic variation in the vomeronasal receptor gene repertoires of inbred mice

ABSTRACT: BACKGROUND: Vomeronasal receptors (VRs), expressed in sensory neurons of the vomeronasal organ, are thought to bind pheromones and mediate innate behaviours. The mouse reference genome has over 360 functional VRs arranged in highly homologous clusters, but the vast majority are of unknown function. Differences in these receptors within and between closely related species of mice are likely to underpin a range of behavioural responses. To investigate these differences, we interrogated the VR gene repertoire from 17 inbred strains of mice using massively parallel sequencing. RESULTS: Approximately half of the 6222 VR genes that we investigated could be successfully resolved, and those that were unambiguously mapped resulted in an extremely accurate dataset. Collectively VRs have over twice the coding sequence variation of the genome average; but we identify striking non-random distribution of these variants within and between genes, clusters, clades and functional classes of VRs. We show that functional VR gene repertoires differ considerably between different Mus subspecies and species, suggesting these receptors may play a role in mediating behavioural adaptations. Finally, we provide evidence that widely-used, highly inbred laboratory-derived strains have a greatly reduced, but not entirely redundant capacity for differential pheromone-mediated behaviours. CONCLUSIONS: Together our results suggest that the unusually variable VR repertoires of mice have a significant role in encoding differences in olfactory-mediated responses and behaviours. Our dataset has expanded over nine fold the known number of mouse VR alleles, and will enable mechanistic analyses into the genetics of innate behavioural differences in mice.     

Recombinant inbred lines for genetic mapping in tomato

A cross between the cultivated tomato Lycopersicon esculentum and a related wild species L. cheesmanii yielded 97 recombinant inbred lines (RILs) which were used to construct a genetic map consisting of 132 molecular markers. Significant deviation from the expected 1:1 ratio between the two homozygous classes was found in 73% of the markers. In 98% of the deviating markers, L. esculentum alleles were present in greater frequency than the L. cheesmanii alleles. For most of the markers with skewed segregation, the direction of the deviation was maintained from F₂ to F₇ generations. The average heterozygosity in the population was 15%. This value is significantly greater than the 1.5% heterozygosity expected for RILs in the F₇ generation. On average, recombination between linked markers was twice as high in the RILs than in the F₂ population used to derive them. The utility of RILs for the mapping of qualitative and quantitative traits is discussed.     

Whole-genome association mapping in elite inbred crop varieties

We have previously shown that linkage disequilibrium (LD) in the elite cultivated barley (Hordeum vulgare) gene pool extends, on average, for <1-5 cM. Based on this information, we have developed a platform for whole genome association studies that comprises a collection of elite lines that we have characterized at 3060 genome-wide single nucleotide polymorphism (SNP) marker loci. Interrogating this data set shows that significant population substructure is present within the elite gene pool and that diversity and LD vary considerably across each of the seven barley chromosomes. However, we also show that a subpopulation comprised of only the two-rowed spring germplasm is less structured and well suited to whole genome association studies without the need for extensive statistical intervention to account for structure. At the current marker density, the two-rowed spring population is suited for fine mapping simple traits that are located outside of the genetic centromeres with a resolution that is sufficient for candidate gene identification by exploiting conservation of synteny with fully sequenced model genomes and the emerging barley physical map.     

An optimal DNA pooling strategy for progressive fine mapping

We present a cost-effective DNA pooling strategy for fine mapping of a single Mendelian gene in controlled crosses. The theoretical argument suggests that it is potentially possible for a single-stage pooling approach to reduce the overall experimental expense considerably by balancing costs for genotyping and sample collection. Further, the genotyping burden can be reduced through multi-stage pooling. Numerical results are provided for practical guidelines. For example, the genotyping effort can be reduced to only a small fraction of that needed for individual genotyping at a small loss of estimation accuracy or at a cost of increasing sample sizes slightly when recombination rates are 0.5% or less. An optimal two-stage pooling scheme can reduce the amount of genotyping to 19.5%, 14.5% and 6.4% of individual genotyping efforts for identifying a gene within 1, 0.5, and 0.1 cM, respectively. Finally, we use a genetic data set for mapping the rice xl(t) gene to demonstrate the feasibility and efficiency of the DNA pooling strategy. Taken together, the results demonstrate that this DNA pooling strategy can greatly reduce the genotyping burden and the overall cost in fine mapping experiments.     

Components of the major urinary protein complex of inbred mice: Separation and peptide mapping

Gel filtration of the nondialyzable fraction of urine from normal inbred mice on Sephadex G-100 yielded three peaks (I, II, and III in order of elution), the relative sizes of which varied with the sex and strain of the mice. Constituents of peak I, the breakthrough peak, included uromucoid (Tamm-Horsfall mucoprotein); peak III was low in nitrogen, rich in carbohydrate, nonprecipitable with trichloroacetic acid, gave no definitive ultraviolet or visible spectrum, and had a sedimentation coefficient of 0.5 S. Peak II contained the electrophoretically distinguishable prealbumins of the major urinary protein (MUP) complex. These components (known as 1, 2, and 3 in order of increasing mobility toward the anode) were separated by chromatography on diethylaminoethyl cellulose. Tryptic peptide mapping indicated that components 1 and 2, a genetic variant shown to be under the control of one genetic locus (the Mup-alocus), differed by a single peptide. Components 1 and 3 had a number of peptides in common plus several peptides unique to each. The peptide map of any given component did not differ between sexes or between the strains investigated.