SNPdetector: a software tool for sensitive and accurate SNP detection

Identification of single nucleotide polymorphisms (SNPs) and mutations is important for the discovery of genetic predisposition to complex diseases. PCR resequencing is the method of choice for de novo SNP discovery. However, manual curation of putative SNPs has been a major bottleneck in the application of this method to high-throughput screening. Therefore it is critical to develop a more sensitive and accurate computational method for automated SNP detection. We developed a software tool, SNPdetector, for automated identification of SNPs and mutations in fluorescence-based resequencing reads. SNPdetector was designed to model the process of human visual inspection and has a very low false positive and false negative rate. We demonstrate the superior performance of SNPdetector in SNP and mutation analysis by comparing its results with those derived by human inspection, PolyPhred (a popular SNP detection tool), and independent genotype assays in three large-scale investigations. The first study identified and validated inter- and intra-subspecies variations in 4,650 traces of 25 inbred mouse strains that belong to either the Mus musculus species or the M. spretus species. Unexpected heterozygosity in CAST/Ei strain was observed in two out of 1,167 mouse SNPs. The second study identified 11,241 candidate SNPs in five ENCODE regions of the human genome covering 2.5 Mb of genomic sequence. Approximately 50% of the candidate SNPs were selected for experimental genotyping; the validation rate exceeded 95%. The third study detected ENU-induced mutations (at 0.04% allele frequency) in 64,896 traces of 1,236 zebra fish. Our analysis of three large and diverse test datasets demonstrated that SNPdetector is an effective tool for genome-scale research and for large-sample clinical studies. SNPdetector runs on Unix/Linux platform and is available publicly (http://lpg.nci.nih.gov).     

A genome-wide set of congenic mouse strains derived from CAST/Ei on a C57BL/6 background

We previously reported the construction of two sets of heterozygous congenic strains spanning the mouse genome. For both sets, C57BL/6J was employed as the background strain while DNA from either DBA/2 or CAST/Ei was introgressed to form the congenic region. We have subsequently bred most of these strains to produce homozygous breeding stocks. Here, we report the characterization of the strain set based on CAST/Ei. CAST/Ei is the most genetically distant strain within the Mus mus species and many trait variations relevant to common diseases have been identified in CAST/Ei mice. Despite breeding difficulties for some congenic regions, presumably due to incompatible allelic variations between CAST/Ei and C57BL/6, the resulting congenic strains cover about 80% of the autosomal chromosomes and will be useful as a resource for the further analysis of quantitative trait loci between the strains.     

Localization of the growth hormone gene to the distal half of mouse chromosome 11

A DNA fragment size variant for the growth hormone gene, Gh, has been identified among inbred strains of mice. The inbred strains SM/J and CAST/Ei carry the less frequent allele Ghb and 11 other strains carry the Gha allele. Segregation analysis of data from two crosses involving SM/J and NZB/BINJ and a cross involving BALB/cJ and CAST/Ei confirmed the assignment of Gh to mouse chromosome 11 and placed the locus 2.6 +/- 1.8 map units distal to Erba (avian erythroblastosis oncogene A), a position consistent with the assignment of the Gh locus to the q22-q24 region of chromosome 17 on the human map. Segregation analysis also refined the location of Sparc (secreted acidic cysteine-rich glycoprotein) on mouse chromosome 11 to a position 16.7 +/- 4.2 map units proximal to Evi-2 (ecotropic viral integration site 2).     

Structure and genetic polymorphism of the mouse KCC1 gene

The KCC1 K-Cl cotransporter is a major regulator of erythroid and non-erythroid cell volume, and the KCC1 gene is a candidate modifier gene for sickle cell disease and other hemoglobinopathies. We have cloned and sequenced the mouse KCC1 (mKCC1) gene, defined its intron-exon junctions, and analyzed (AC)/(TG) intragenic polymorphisms. A highly polymorphic (AC) repeat of mKCC1 intron 1 was characterized in musculus strains, and used to prove lack of linkage between the mKCC1 gene and the rol (resistant to osmotic lysis) locus. The intron 1 (AC) repeat in CAST/Ei and SPRET/Ei was not only more divergent in length but also underwent additional sequence variation. A dimorphic (TG) repeat in intron 2 distinguished CAST/Ei from other strains, and an intron 17 B1 Alu-like SINE present in all musculus strains was found to be absent from intron 17 in SPRET/Ei. These and additional described strain-specific polymorphisms will be useful mapping and genetic tools in the study of mouse models of sickle cell disease.     

New murine polymorphisms detected by random amplified polymorphic DNA (RAPD) PCR and mapped by use of recombinant inbred strains

Oligonucleotide primers of random sequence that were 12 bases in length, 58% in GC content, and lacking internal palindromes were designed. By random amplified polymorphic DNA (RAPD) PCR, these primers were used to survey for DNA variations between the progenitors of the mouse AXB and BXA recombinant inbred sets (A/J and C57BL/6J). We identified 17 DNA variants detected by 10 primers. Map positions for these variants were determined by comparing their strain distribution patterns in the AXB, BXA recombinant inbred sets with strain distribution patterns of previously published loci. When necessary, BXD and NXSM recombinant inbred sets were also used. These 17 new loci mapped to 12 chromosomes. The 10 primers were also used to survey 20 inbred mouse strains including the progenitors of other recombinant inbred sets and four mouse strains recently inbred from the wild (CAST/Ei, MOLF/Ei, PERA/Ei, and SPRET/Ei).     

Genetic control of HDL levels and composition in an interspecific mouse cross (CAST/Ei x C57BL/6J)

Strain CAST/Ei (CAST) mice exhibit unusually low levels of high density lipoproteins (HDL) as compared with most other strains of mice, including C57BL/6J (B6). This appears to be due in part to a functional deficiency of lecithin:cholesterol acyltransferase (LCAT). LCAT mRNA expression in CAST mice is normal, but the mice exhibit several characteristics consistent with functional deficiency. First, the activity and mass of LCAT in plasma and in HDL of CAST mice were reduced significantly. Second, the HDL of CAST mice were relatively poor in phospholipids and cholesteryl esters, but rich in free cholesterol and apolipoprotein A-I (apoA-I). Third, the adrenals of CAST mice were depleted of cholesteryl esters, a phenotype similar to that observed in LCAT- and acyl-CoA:cholesterol acyltransferase-deficient mice. Fourth, in common with LCAT-deficient mice, CAST mice contained triglyceride-rich lipoproteins with "panhandle"-like protrusions. To examine the genetic bases of these differences, we studied HDL lipid levels in an intercross between strain CAST and the common laboratory strain B6 on a low fat, chow diet as well as a high fat, atherogenic diet. HDL levels exhibited complex inheritance, as 12 quantitative trait loci with significant or suggestive likelihood of observed data scores were identified. Several of the loci occurred over plausible candidate genes and these were investigated.The results indicate that the functional LCAT deficiency is unlikely to be due to variations of the LCAT gene. Our results suggest that novel genes are likely to be important in the control of HDL metabolism, and they provide evidence of genetic factors influencing the interaction of LCAT with HDL.     

Macronutrient diet selection in thirteen mouse strains

The strain distribution for macronutrient diet selection was described in 13 mouse strains (AKR/J, NZB/B1NJ, C57BL/6J, C57BL/6ByJ, DBA/2J, SPRET/Ei, CD-1, SJL/J, SWR/J, 129/J, BALB/cByJ, CAST/Ei, and A/J) with the use of a self-selection protocol in which separate carbohydrate, fat, and protein diets were simultaneously available for 26-30 days. Relative to carbohydrate, nine strains consumed significantly more calories from the fat diet; two strains consumed more calories from carbohydrate than from fat (BALB/cByJ, CAST/Ei). Diet selection by SWR/J mice was variable over time, resulting in a lack of preference. One strain (A/J) failed to adapt to the diet paradigm due to inadequate protein intake. Comparisons of proportional fat intake across strains revealed that fat selection/consumption ranged from 26 to 83% of total energy. AKR/J, NZB/B1NJ, and C67BL/6J mice self-selected the highest proportion of dietary fat, whereas the CAST/Ei and BALB/cByJ strains chose the lowest. Finally, epididymal fat depot weight was correlated with fat consumption. There were significant positive correlations in AKR/J and C57BL/6J mice, which are highly sensitive to dietary obesity. However, absolute fat intake was inversely correlated with epididymal fat in two of the lean strains: SWR/J and CAST/Ei. We hypothesize that the SWR/J and CAST/Ei strains are highly sensitive to a negative feedback signal generated by increasing body fat, but the AKR/J and C67BL/6J mice are not. The variation in dietary fat selection across inbred strains provides a tool for dissecting the complex genetics of this trait.     

Genetic basis of murine responses to hyperoxia-induced lung injury

To evaluate the effect of genetic background on oxygen (O₂) toxicity, nine genetically diverse mouse strains (129/SvIm, A/J, BALB/cJ, BTBR+(T)/tf/tf, CAST/Ei, C3H/HeJ, C57BL/6J, DBA/2J, and FVB/NJ) were exposed to more than 99% O₂ for 72 h. Immediately following the hyperoxic challenge, the mouse strains demonstrated distinct pathophysiologic responses. The BALB/cJ and CAST/Ei strains, which were the only strains to demonstrate mortality from the hyperoxic challenges, were also the only strains to display significant neutrophil infiltration into their lower respiratory tract. In addition, the O₂-challenged BALB/cJ and CAST/Ei mice were among six strains (A/J, BALB/cJ, CAST/Ei, BTBR+(T)/tf/tf, DBA/2J, and C3H/HeJ) that had significantly increased interleukin 6 concentrations in the whole lung lavage fluid and were among all but one strain that had large increases in lung permeability compared with air-exposed controls. In contrast, the DBA/2J strain was the only strain not to have any significant alterations in lung permeability following hyperoxic challenge. The expression of the extracellular matrix proteins, including collagens I, III, and IV, fibronectin I, and tenascin C, also varied markedly among the mouse strains, as did the activities of total superoxide dismutase (SOD) and manganese-SOD (Mn-SOD or SOD2). These data suggest that the response to O₂ depends, in part, on the genetic background and that some of the strains analyzed can be used to identify specific loci and genes underlying the response to O₂.     

Multi-phenotype behavioral characterization of inbred strains derived from wild stocks of Mus musculus

Many aspects of mouse behavior have been studied by using only a relatively small sample of available laboratory strains. These laboratory mice were derived from the so-called ``fancy mouse' and in most cases underwent extensive domestication before inbreeding. Thus, the behavioral repertoire of the laboratory mouse may be very different from that exhibited by stocks that have not been deliberately domesticated. Another inherent problem in analyzing mouse behavior is that genetic diversity is limited among currently available strains. In this respect, the use of strains that are derived from a variety of wild mice should provide a means to identifying novel behavioral phenotypes. We have investigated several behavioral phenotypes, using females of a number of mouse strains derived from wild mice of different subspecies, BFM/2, NJL, BLG2, HMI, CAST/Ei, KJR, SWN and MSM; a strain derived from fancy mice, JF1; and two laboratory strains, C57BL/6 and DBA/1. In this report, tests for locomotor activity, light-dark transitions, passive and active avoidance, and nociception were conducted. The results show great diversity of behavioral patterns between strains in contrast to less within-strain variability. We also found that two strains, KJR and SWN, both have good learning ability, whereas BLG2 mice exhibit impairment in both passive and active avoidance learning. Received: 11 January 2000 / Accepted: 27 March 2000     

Linkage of the mouse LDL receptor gene on chromosome 9

We identified restriction fragment length variants of the mouse low-density lipoprotein receptor gene and used these to map the gene, designated Ldlr, to the proximal region of chromosome 9. An interspecific backcross between strains MEV and CAST/Ei yielded the following gene order and distances in centimorgans: Ldlr-18.8 +/- 5.6-Apoa-4-7.3 +/- 3.5-Mpi-1-10.2 +/- 3.9-Emv-3 or dilute. Analysis of recombinant inbred strains also indicated that Ldlr is tightly linked to a previously unmapped retroviral marker, Xmmv-67.     

Phylogenetic relationships among laboratory and wild-origin Mus musculus strains on the basis of genomic DNA RFLPs

Genetic distance measures between the laboratory mouse strains C57BL/6J and RF/J and the wild-origin Mus musculus mouse strains CAST/Ei, MOLF/Ei, POSCH I, and CZECH II were estimated by allelic patterns revealed by RFLP analysis. These results suggest phylogenetic relationships indicating that the mouse strains related to the subspecies M.m. domesticus (RF/J, POSCH I and C57BL/6J) are more closely related to the CAST/Ei strain (derived from M.m. castaneus) than to the strains CZECH II (M.m. musculus) and MOLF/Ei (M.m. molossinus). Furthermore, the hybrid strain C57BL/6J is more closely related to POSCH I (M.m. poschiavinus) than to RF/J as calculated by the method distance measures of Cavalli-Sforza and Edwards (Evolution 21,550, 1967), Nei's minimum (Am. Natural. 106,283, 1972) and unbiased minimum (Genetics 89,583, 1978), Edwards (Biometrics 27,873, 1971; Genetic Distance, p. 41, 1974) and Rogers modified (1986).     

Effects of natural selection on interpopulation divergence at polymorphic sites in human protein-coding Loci

To develop new strategies for searching for genetic associations with complex human diseases, we analyzed 2784 single-nucleotide polymorphisms (SNPs) in 396 protein-coding genes involved in biological processes relevant to cancer and other complex diseases, with respect to gene diversity within samples of individuals representing the three major historic human populations (African, European, and Asian) and with respect to interpopulation genetic distance. Reduced levels of both intrapopulation gene diversity and interpopulation genetic distance were seen in the case of SNPs located within the 5'-UTR and at nonsynonymous SNPs, causing radical changes to protein structure. Reduction of gene diversity at SNP loci in these categories was evidence of purifying selection acting at these sites, which in turn causes a reduction in interpopulation divergence. By contrast, a small number of SNP sites in these categories revealed unusually high genetic distances between the two most diverged populations (African and Asian); these loci may have historically been subject to divergent selection pressures.     

cDNA sequence of five mouse guanine deaminase (Gda) alleles and mapping to mouse chromosome 19.

Guanine deaminase catalyses the conversion of guanine to xanthine and ammonia, thereby irreversibly removing the guanine base from the pool of guanine-containing metabolites. We have identified five alleles at the mouse guanine deaminase locus by cDNA sequencing. These alleles were defined by single-nucleotide polymorphisms at a total of 19 positions. For each allele the representative strains are as follows: Gda(a), C57BL/6J and DBA/2J; Gda(b), A/J; Gda(c), MOLF/Ei; Gda(d), CAST/Ei; and Gda(e), SPRET-1. The only codon change resulting in an amino acid substitution was found at nucleotide 523, where GAT was replaced by AAT in Mus spretus resulting in the deduced substitution of Asp-174 by Asn. The single-nucleotide difference between the a and b alleles was also typed by allele-specific oligonucleotide amplification for 17 common strains of Mus musculus susbp. musculus. By typing the AxB and BxA recombinant inbred (RI) strain sets, Gda was mapped to mouse chromosome 19, a region syntenic with human chromosome 9q11-q22.     

Valproic acid increases expression of methylenetetrahydrofolate reductase (MTHFR) and induces lower teratogenicity in MTHFR deficiency

Valproate (VPA) treatment in pregnancy leads to congenital anomalies, possibly by disrupting folate or homocysteine metabolism. Since methylenetetrahydrofolate reductase (MTHFR) is a key enzyme of folate interconversion and homocysteine metabolism, we addressed the possibility that VPA might have different teratogenicity in Mthfr(+/+) and Mthfr(+/-) mice and that VPA might interfere with folate metabolism through MTHFR modulation. Mthfr(+/+) and Mthfr(+/-) pregnant mice were injected with VPA on gestational day 8.5; resorption rates and occurrence of neural tube defects (NTDs) were examined on gestational day 14.5. We also examined the effects of VPA on MTHFR expression in HepG2 cells and on MTHFR activity and homocysteine levels in mice. Mthfr(+/+) mice had increased resorption rates (36%) after VPA treatment, compared to saline treatment (10%), whereas resorption rates were similar in Mthfr(+/-) mice with the two treatments (25-27%). NTDs were only observed in one group (VPA-treated Mthfr(+/+)). In HepG2 cells, VPA increased MTHFR promoter activity and MTHFR mRNA and protein (2.5- and 3.7-fold, respectively). Consistent with cellular MTHFR upregulation by VPA, brain MTHFR enzyme activity was increased and plasma homocysteine was decreased in VPA-treated pregnant mice compared to saline-treated animals. These results underscore the importance of folate interconversion in VPA-induced teratogenicity, since VPA increases MTHFR expression and has lower teratogenic potential in MTHFR deficiency.     

Functional and genetic survey of all known single-nucleotide polymorphisms within the human deoxyribonuclease I gene in wide-ranging ethnic groups

The single-nucleotide polymorphisms (SNPs) in the human DNase I gene (DNASE1) might be involved in susceptibility to some common diseases; however, only limited population data are available. Further, the effects of these SNPs on in vivo DNase I activity remain unknown. The genotype and haplotype of all the SNPs in DNASE1 were determined in 3 ethnic groups including 14 populations using newly developed methods. Together with our previous data on the nonsynonymous SNPs, two major haplotypes based on the five exonic SNPs were identified; genetic diversity in the Asian population was low. Among 10 SNPs, other than exonic SNPs in the gene, only 3 were polymorphic among all the populations. Haplotype distribution, based on all the polymorphic SNPs, was clarified to be generally varied in an ethnic-dependent manner. Thus, the genetic aspects of DNASE1 with regard to all the SNPs in wide-ranging ethnic groups could be first demonstrated. Further, there was no correlation of all the polymorphic SNPs other than nonsynonymous ones with serum DNase I activity levels. Polymorphic SNPs other than the exonic SNPs might not be directly related to common diseases through alterations in in vivo levels of the activity.     

Molecular evolution of cadherin-related neuronal receptor/protocadherin(alpha) (CNR/Pcdh(alpha)) gene cluster in Mus musculus subspecies

The mouse cadherin-related neuronal receptor/protocadherin (CNR/Pcdh) gene clusters are located on chromosome 18. We sequenced single-nucleotide polymorphisms (SNPs) of the CNR/Pcdh(alpha)-coding region among 12 wild-derived and four laboratory strains; these included the four major subspecies groups of Mus musculus: domesticus, musculus, castaneus, and bactrianus. We detected 883 coding SNPs (cSNPs) in the CNR/Pcdh(alpha) variable exons and three in the constant exons. Among all the cSNPs, 586 synonymous (silent) and 297 nonsynonymous (amino acid exchanged) substitutions were found; therefore, the K(a)/K(s) ratio (nonsynonymous substitutions per synonymous substitution) was 0.51. The synonymous cSNPs were relatively concentrated in the first and fifth extracellular cadherin domain-encoding regions (ECs) of CNR/Pcdh(alpha). These regions have high nucleotide homology among the CNR/Pcdh(alpha) paralogs, suggesting that gene conversion events in synonymous and homologous regions of the CNR/Pcdh(alpha) cluster are related to the generation of cSNPs. A phylogenetic analysis revealed gene conversion events in the EC1 and EC5 regions. Assuming that the common sequences between rat and mouse are ancestral, the GC content of the third codon position has increased in the EC1 and EC5 regions, although biased substitutions from GC to AT were detected in all the codon positions. In addition, nonsynonymous substitutions were extremely high (11 of 13, K(a)/K(s) ratio 5.5) in the laboratory mouse strains. The artificial environment of laboratory mice may allow positive selection for nonsynonymous amino acid variations in CNR/Pcdh(alpha) during inbreeding. In this study, we analyzed the direction of cSNP generation, and concluded that subspecies-specific nucleotide substitutions and region-restricted gene conversion events may have contributed to the generation of genetic variations in the CNR/Pcdh genes within and between species.