Drs. Gauguier and van Luijtelaar contributed equally to this study.
Chromosomal Mapping of Genetic Loci Controlling Absence Epilepsy Phenotypes in the WAG/Rij Rat
Version of Record online: 21 JUL 2004
Volume 45, Issue 8, pages 908–915, August 2004
How to Cite
Gauguier, D., Luijtelaar, G. v., Bihoreau, M. T., Wilder, S. P., Godfrey, R. F., Vossen, J., Coenen, A. and Cox, R. D. (2004), Chromosomal Mapping of Genetic Loci Controlling Absence Epilepsy Phenotypes in the WAG/Rij Rat. Epilepsia, 45: 908–915. doi: 10.1111/j.0013-9580.2004.13104.x
- Issue online: 21 JUL 2004
- Version of Record online: 21 JUL 2004
- Accepted March 24, 2004.
- WAG/Rij rat;
- Quantitative trait locus;
- Absence epilepsy;
- Spike–wave discharges;
Summary: Purpose: The WAG/Rij rat is among the most appropriate models for the study of spontaneous childhood absence epilepsy, without complex neurologic disorders that are associated with some mouse models for absence epilepsy. Previous studies have allowed the identification of distinct types of spike–wave discharges (SWDs) characterizing seizures in this strain. The purpose of this study was to investigate the genetic basis of electroencephalographic (EEG) properties of SWDs.
Methods: An intercross was derived from WAG/Rij and ACI inbred strains that are known to differ substantially in the number of SWDs. Phenotypic analyses based on 23-h EEG recording in all progenies allowed the quantification of type I and type II SWD phenotypes. A genome-wide scan was performed with 145 microsatellite markers, which were used to test for evidence of genetic linkage to SWD quantitative phenotypes.
Results: We were able to map quantitative trait loci independently, controlling type I and type II SWD variables to rat chromosomes 5 and 9. Strongest linkages were obtained for D5Mgh15 and total duration of type II SWD (lod, 3.64) and for D9Rat103 and the average duration of type I SWD (lod, 3.91). These loci were denoted T2swd/wag and T1swd/wag, respectively.
Conclusions: The independent genetic control of type I and type II SWDs underlines the complexity of the molecular mechanisms participating in SWDs. The identification of these genetic loci represents an important step in our fundamental knowledge of the architecture of SWDs and may provide new insights for resolving the genetic heterogeneity of absence epilepsy.
Idiopathic generalized epilepsy (IGE) is a group of complex disorders, which includes childhood absence epilepsy (CAE). The difference from other forms of absence epilepsy such as juvenile absence epilepsy (JAE) is the age at onset and the frequency of occurrence. CAE is characterized by an early age at onset and up to several hundreds of attacks per day, whereas JAE starts around puberty with absences occurring less frequently than one per day. CAE or pyknolepsy is the archetypical epileptic syndrome of typical absence seizures (1). It is characterized by typical bilateral generalized spike–wave discharges (SWDs) on the EEG activity and a concomitant impairment of consciousness (absence). Absences are brief (2–5 s) or long (15–30 s) and have a sudden onset and termination. The background EEG activity is usually normal, although paroxysmal activity may occur (2). Estimations of incidence range from 0.7 to 4.6 per 100,000 in the general population and from 6 to 8 per 100,000 in children and adolescents up to age 15 years (3). The proportion of patients with typical absences among people with epilepsy is ∼3%, and the prevalence of typical absences among children with epilepsies is ∼10% (1). Absences usually respond well to ethosuximide (ESM) or sodium valproate (VPA) and remit within 2–5 years from onset (2).
Both the high concordance (>85%) of IGE in monozygotic twins and results from segregation studies support the existence of a strong genetic component in the etiology of this disease (4,5). Results from linkage and association studies in humans have demonstrated the existence of multiple genes involved in various forms of the disease and highlighted the complexity of genetic investigations in human epilepsy (6). Evidence of genetic linkage to human 8q24 was found in a particular type of CAE (evidence of absence epilepsy highly characteristic SWD plus tonic–clonic generalized seizures) and replicated in two independent studies (7,8). An association between CAE and genetic variants in the γ-aminobutyric acid type A–receptor subunit β3 (GABRB3) located on human 15q11-q13 was found, suggesting that these alleles may be either directly or indirectly involved in the etiology of CAE (9). Finally, several lines or research suggest that Ca2+ channel deficits might be involved in increased seizure sensitivity (10). However, the identification of genes involved in this pathology is hampered mainly by the lack of large, well-documented families with a history of EEG-confirmed epilepsy, the heterogeneous pattern of EEGs characterizing absence epilepsy, and an incomplete phenotypic expression in patients.
Our knowledge of the pathophysiology and etiology of CAE has recently progressed because of the availability of inbred rodent strains and selection lines spontaneously exhibiting key features of the human disease. Genetic studies in murine models have highlighted the complex regulation of SWDs, the EEG characteristic of absence epilepsy (11), and highlighted the importance of Ca2+ channels in regulating the expression of generalized cortical SWDs (12). Rat models of human epilepsy represent essential tools with which to perform accurately extensive phenotypic investigations, including in particular prolonged EEG measurements that remain technically challenging in mice.
The Wistar Albino Glaxo rat (WAG/Rij) and the genetic absence epilepsy rat from Strasbourg (GAERS) spontaneously show hundreds of SWDs per day and are among the most appropriate models for typical absence epilepsy (13–16). The WAG/Rij rat exhibits, besides this generalized, bilateral symmetrical SWD (type I), a more local occipital–parietal type II SWD (17,18). A wide variability of SWD occurrence was observed in progenies of WAG/Rij × ACI F2 and reciprocal backcross and F344 × BN rat crosses, indicating the involvement of several modulating genes in the control of SWD characteristics (19–22).
The aim of the present study was to dissect accurately absence epilepsy in SWD quantitative subphenotypes in F2 progenies derived from WAG/Rij and ACI rats and locate quantitative trait loci (QTLs) controlling these traits. We were able to identify two QTLs independently controlling type I and type II SWD variables that differ in many aspects in the WAG/Rij strain. The identification of genetic loci controlling SWD subphenotypes represents an important step in the resolution of the genetic heterogeneity of absence epilepsy.
MATERIAL AND METHODS
Animals and experimental cross
All rats were maintained under a 12 h/12 h light–dark regimen with white lights off at 9 p.m. Food and water were at all times provided ad libitum. Animals were maintained in accordance with the principles of Laboratory Animal Care and institutional guidelines. All procedures were carried out in accordance with our institutional guidelines.
Parental inbred strains (ACI, agouti, and WAG/Rij, albino) were bred and maintained in the Department of Biological Psychology of Nijmegen University. An F1 population was initially produced by breeding ACI males and WAG/Rij females. An F2 population was subsequently produced by breeding 16 F1 pairs. In total, 141 progeny were born within a 4-day interval. One litter with an excess of males was discarded from the experiment, and the size of all other litters was randomly adjusted to a maximum of five males and five females to reduce environmental variation. The final number of subjects was 59 males and 59 females, and both genotype and phenotype analyses were successful in 87 animals (40 females and 47 males). At PN 14–17, all F2 rats were numbered, and tail clips were collected for DNA extraction. Rats were weaned at PN 30 and maintained in a standard macrolon cage with a maximum of three littermates of the same sex. Five- to 6-month-old rats were prepared for EEG recordings.
A tripolar EEG electrode set (MS 333:2-A; Plastic One, Roanoke, VI, U.S.A.) was permanently implanted under isoflurane anesthesia. One active electrode was aimed at the frontal cortex [coordinates with the skull surface flat and bregma zero-zero: A: 2.0; L, −3.5 referring to Paxinos and Watson (23)], the second one at the parietal/occipital cortex (A: −6.0, L, −4.0); the reference electrode was placed above the cerebellum. After surgery they were housed separately and were allowed to recover for ≥2 weeks. Next they were familiarized with the EEG recording situation for ≥12 hours. The EEGs were subsequently recorded for 23 h in unrestrained animals with the data-acquisition system WinDaq. EEG filter settings ensured that the EEG between 1 and 100 Hz was allowed to pass. EEGs were sampled at 200 Hz for subsequent storage for off-line analyses. All recordings were obtained within a 4-week period.
Procedures carried out in the F2 cohort resulted in the acquisition of >2,000 h of EEG recordings. The 23-h recording collected in each hybrid was off-line analyzed with a custom-made software package. This system was evaluated by comparing on-line detection of aberrant EEG phenomena of nine rats (2 h each) to the consensus analysis of two experts. A consensus was reached in a total of 405 SWDs. The automatic system detected 97% of the phenomena correctly. Data were first analyzed with this routine and subsequently verified by visual observation. The onset times (detection of the first spike of a train of SWDs) and offset time (last spike of a train of SWDs) of SWD events were adjusted if necessary. Criteria for SWD type classification were as previously published and included an analysis of the topographic distribution of type I and II SWDs (17,18) (Fig. 1). Number, average, and total duration of (trains of) SWDs type I and II were determined over the whole recording period. The SWDs type I and II differ in polarity of the spikes, the mean duration, and the mean frequency, as described in WAG/Rij rats. The scoring of the EEGs was done independently from the genotyping. Considering the high performance of automated SWD phenotype acquisition and subsequent visual checks by EEG experts, SWD phenotyping was highly reliable.
For the genome-wide search of genetic loci controlling SWD, 145 microsatellite markers exhibiting evidence of allele variations between the WAG/Rij and ACI rat strains were selected in our database (http://www.well.ox.ac.uk/rat_mapping_resources) and used to determine the genotypes of the F2 hybrids. All selected markers have already been genetically mapped in various rat crosses (24–26) and/or localized in the rat genome by using the T55 rat radiation hybrid panel (27). Markers used in this study were chosen to cover the 21 chromosomes of the rat genome with an average spacing of ∼10 cM between loci. Genotypes were determined after polymerase chain reaction (PCR) amplification and gel electrophoresis of PCR products, as previously described (24). Oligonucleotides were synthesized commercially by Genosys Biotechnologies (Pampisford, U.K.).
Construction of the genetic maps
Before linkage analysis, JoinMap version 2.0 (28) was used to calculate single-factor segregation ratios for each marker in a linkage group. The genotypes were checked for all individuals at markers that were exhibiting significantly distorted segregation. Subsequently, initial maps were created, double-recombination events were identified with JoinMap, and the data corresponding to these were verified. The JoinMap module for genotype checking calculates for all loci and for all individuals the probability of obtaining the present genotype, conditional on both the genotypes at the two flanking loci and on map distances. We considered unexplained genotypes to be those having a threshold of >3 for the test statistic of log10(1/p).
Before QTL analysis, all phenotypes were corrected for sex by regression. Standardized residuals were used in all subsequent analyses. Nonparametric linkage analysis for quantitative traits was performed by an analysis of variance (ANOVA) test followed by a permutation test (n = 10,000) (30,31) to evaluate the threshold of significance for each pair of genetic–phenotypic markers, as determined by Lander and Kruglyak (49). Correlations between marker genotypes and phenotypes were calculated by using the SPSS version 11.0 package.
All (WAG/Rij × ACI) F2 rats showed SWD type I, and ∼50% (44 of 87) showed SWD type II. Both types of SWDs appeared qualitatively similar to what was commonly seen in WAG/Rij rats. The major characteristics of the SWDs are presented in Table 1. For any of the variables, neither gender differences nor nest effects were statistically significant. Of 87 F2 animals, 23 (26%) had less than 1 SWD per hour (the average number of SWDs in the ACI strain), and 10 (11%) showed a number of SWDs in the range (mean ± 1 SEM) of the WAG/Rij parental strain. The mean number of SWDs per hour was 5.88, lower than could be expected from the mean of the two parental strains taken together. Table 2 shows the correlations between number, mean, and total duration of type I and type II SWDs in males and females of the WAG/Rij × ACI cross. Most striking and significant were the negative correlations (Pearson) between the characteristics of type I and type II SWDs.
|t Value||Males (n = 47)||Females (n = 40)|
|Number of SWDs, type I||0.86, df 85||114.4 ± 12.8||116.4 ± 20.7|
|Mean duration of SWDs, type I||0.57, df 52||2.79 ± 0.22||3.03 ± 0.38|
|Total duration of SWDs, type I||0.12, df 85||465.5 ± 76.4||450.6 ± 99.8|
|Number of SWDs, type II||1.24, df 85||14.1 ± 5.7||27.8 ± 10.2|
|Mean duration of SWDs, type II||0.65, df 42||2.18 ± 0.21||2.01 ± 0.16|
|Total duration of SWDs, type II||0.89, df 85||41.7 ± 20.1||71.2 ± 27.6|
|SWD: Type and characteristic||Number of SWDs, type I||Mean duration of SWDs, type I||Total duration of SWDs, type I|
|Number of SWDs, type II||−0.30||−0.22||−0.44|
|p = 0.005||p = 0.038||p = 0.0001|
|Mean duration of SWDs, type II||−0.28||−0.21||−0.41|
|p = 0.009||p = 0.0001|
|Total duration of SWDs, type II||−0.13||−0.059||−0.174|
A genetic map was constructed with genotype data obtained from the 145 markers typed in the F2 hybrids. Marker order and genetic distances between marker loci were consistent with published maps of the rat genome (25–27). Based on a total genetic length of the rat genome of 1,736 cM (25), the average spacing between adjacent markers was 12 cM. The largest gaps in the linkage map derived from the F2 cross were in the telomeric region of chromosome 2 (15 cM) and in the central region of chromosome 14 (20 cM).
Evidence of significant genetic linkage to SWD subphenotypes was obtained with markers mapped to 20-cM regions of rat chromosomes 5 and 9 (Figs. 2 and 3). Strongest linkages were obtained for D5Mgh15 and the total duration of type II SWD (LOD, 3.64; p = 0.0002) and for D9Rat103 and the average duration of type I SWD (lod, 3.91; p = 0.00012). These loci were denoted T2swd/wag and T1swd/wag, respectively. Because of the relatively small number of phenotypes considered for QTL analysis in the (WAG × ACI) F2 hybrids and the existence of strong correlations between some of the traits (see Table 2), corrections for multiple testing had little effect on QTL significance (data not shown). Each locus accounts for ≤15% of the total phenotypic variance of each linked trait in the cross.
At the locus T2swd/wag, homozygous genotypes WAG/WAG at marker D5Mgh15 contribute to increased numbers, mean, and total duration of SWDs (maximum F value: 3.56; p = 0.033) in the F2 animals, whereas at marker D9Rat103 (locus T1swd/wag), they have opposite effects on mean and total duration of type I (maximum F value: 4.19; p = 0.0085) (Table 3A and B). Furthermore, when SWDs were dissected into subphenotypes, WAG alleles at marker locus D5Mgh15 (T2swd/wag) specifically contributed to a strongly significant increase in numbers, mean, and total duration of type II SWDs (maximum F value: 9.83; p = 0.00016) and are not correlated with changes in any of the type I SWD phenotypes in the cross (see Table 3A). A similar effect of WAG alleles on type II SWD phenotypes was observed at marker locus D9Rat103 (T1swd/wag), but to a lesser extent, and phenotypic differences between animals carrying each of the three genotypes were not statistically significant (see Table 3B). In contrast, homozygous ACI/ACI animals at the locus show a significant increase in the mean and total duration of type I SWDs (maximum F value: 6.09; p = 0.00091) (see Table 3B).
|A: D5Mgh15 Phenotype||Genotype|
|Lod||Lod threshold||ACI/ACI (26)||WAG/ACI (39)||WAG/WAG (14)||F||p Value|
|Type I SWDs||Number||0.2||3.1||90 ± 20||97 ± 17||99 ± 28||NS|
|Duration||0.6||3.1||347 ± 114||411 ± 93||453 ± 132||NS|
|Average||0.5||3.1||2.55 ± 0.33||2.72 ± 0.30||3.12 ± 0.70||NS|
|Type II SWDs||Number||2.9||3.1||8.8 ± 6.0||10.8 ± 4.1||65.0 ± 26.0||7.5||0.001|
|Duration||3.8||2.9||17.8 ± 13.9||23.1 ± 8.5||208 ± 85||9.8||0.0002|
|Average||2.2||3.1||0.70 ± 0.16||1.01 ± 0.18||1.91 ± 0.43||5.2||0.008|
|Types I & II||Number||1||3||99 ± 19||108 ± 16||164 ± 24||NS|
|SWDs||Duration||0.6||3.1||365 ± 113||434 ± 92||661 ± 109||NS|
|combined||Average||1.5||2.8||3.25 ± 0.39||3.72 ± 0.30||5.04 ± 0.67||3.6||0.03|
|B: D9RAT103 Phenotype||Genotype|
|Lod||Lod threshold||ACI/ACI (21)||WAG/ACI (40)||WAG/WAG (18)||F||p Value|
|Type I SWDs||Number||1.4||3.6||137 ± 24||72 ± 14||102 ± 27||NS|
|Duration||1.8||2.8||680 ± 146||239 ± 64||455 ± 156||3.2||0.03|
|Average||3.9||3||4.15 ± 0.48||2.11 ± 0.22||2.58 ± 0.47||6.1||0.0009|
|Type II SWDs||Number||0.6||2.9||4.2 ± 2.2||23.4 ± 9.7||29.9 ± 13.3||NS|
|Duration||0.5||2.8||14.5 ± 8.9||58.4 ± 26.5||91 ± 50||NS|
|Average||0.4||2.9||0.88 ± 0.30||1.07 ± 0.17||1.33 ± 0.32||NS|
|Types I & II||Number||0.7||2.9||141 ± 24||96 ± 15||132 ± 24||NS|
|SWDs||Duration||1.6||2.9||694 ± 146||297 ± 64||546 ± 149||2.9||0.04|
|combined||Average||2.7||2.9||5.03 ± 0.60||3.18 ± 0.26||3.91 ± 0.40||4.2||0.009|
We successfully characterized two QTLs that independently control SWD variables in a WAG/Rij × ACI cross. Based on specific SWD features exhibited by the WAG/Rij rat, we were able to dissect the SWD phenotype into different quantitative EEG components and demonstrated that the QTL T2swd/wag in rat chromosome 5 controls type II SWD, whereas type I SWDs are under the control of the locus T1swd/wag, mapped to rat chromosome 9. The independent genetic control of types I and II SWDs addresses important concepts in the complex etiology and pathophysiology of absence epilepsy.
A polygenic control of SWD variables was recently reported in an F2 cross derived from the GAERS and Brown Norway (BN) rat strains (32). Although both GAERS and WAG/Rij inbred rats originate from outbred Wistar stocks, we mapped SWD-related QTLs to different chromosomes in the GAERS × BN and WAG/Rij × ACI crosses, suggesting that distinct collections of gene variants regulating SWDs have been selectively fixed in the GAERS and WAG/Rij rats. Differences in the experimental design of the two studies, including phenotype procedures performed in the F2 cohorts, as well as genomic differences of the “SWD-resistant” strains chosen for the experimental crosses, also may account for discordant QTL mapping. Results obtained in the two crosses are complementary and initiate the search for gene variants involved in the development of SWDs in rats.
The use of rat models was a key component in our genetic study, as it allowed us to carry out accurate 23-h EEG recordings in F2 hybrids and subsequently to quantify SWD intermediate phenotypes. Despite the relatively modest QTL effect, which may be due to the low number of F2 rats used and the requirement for extensive phenotype procedures, we were able to determine the role of genetic loci on the control of different types of SWDs. Interestingly, significant genetic linkage was specifically found to the number and total duration of SWD type II, and the mean duration, to a lesser extent. Although the description of this type of EEG oscillation in WAG/Rij rats originates from 1986 (17), virtually nothing is known about type II SWDs. All WAG/Rij rats develop type I SWDs, but only ∼60% of them exhibit type II SWD. Furthermore, the intraspike frequency (the number of individual spikes per second within a train of spike–waves), number and mean duration of type II SWDs are lower than those of type I SWDs (17). Both type I and II SWDs can be aggravated by the GABA-reuptake inhibitor tiagabine (TGB) (33). It was recently reported that type II SWD is, again in contrast to type I SWD, a local EEG oscillation with a parietal/occipital cortical distribution (18). Moreover, type II SWD occurs at a cortical region at which the wave of a type I SWD, the inhibitory part of a spike–wave complex, is maximally expressed. Most important, dopaminergic (DA) antagonists enhance the number of SWDs type I (34–36), whereas DA agonists, such as cocaine and apomorphine, not only reduce type I SWDs, but they also enhance type II SWDs (18,37). The detection of a specific linkage to type II SWD phenotypes suggests an effect of gene(s) at the locus T2swd/wag on the control of excitability of the central nervous system at the parietal/occipital cortex.
Linkage between T1swd/wag and the mean and total duration of SWDs appears to account for the specific linkage to type I SWD phenotypes. Unexpectedly, rats carrying the ACI/ACI genotype at the locus T1swd/wag have a long mean duration of SWDs compared with rats carrying the other two genotypes. However, absence epilepsy in laboratory rats is a very common phenomenon that can be observed in many outbred rat colonies, such as Wistar and Long-Evans, as well as inbred strains such as the G, B, and Brown Norway rats, especially when rats become old (38–42). We have shown that BN alleles can be associated with an increased frequency and amplitude of SWDs in a GAERS × BN F2 cross (32). Although young ACI rats are devoid of SWDs, or show only an incidental SWD (19), 16-month-old ACI rats exhibit a high incidence of SWDs, though it remains much lower than in 6-month-old WAG/Rij rats (36). Detailed quantitative EEG studies showed that 6-month-old ACI rats do have 0.86 SWDs per hour, which is only a fraction of what is observed in age-matched WAG/Rij rats (14.1 SWDs per hour) (36).
Although the WAG/Rij and ACI strains markedly differ in the age at which SWDs are fully expressed in the EEG, genetic linkage to the number of type I SWDs was not detected in the cross. We found evidence of genetic linkage to the mean and total duration of type I SWD, suggesting a differential effect of genes on mechanisms triggering SWDs. Interestingly, Peeters et al. (20,21) described differential genetic models for the average duration rather than the number of SWDs (type I). Additive, dominance, and maternal influences were found for the average duration of SWDs, whereas only dominance was found for the number of SWDs. In our cross, the average duration of type I SWD was reduced in both homozygous WAG/WAG and heterozygous rats at the QTL T1swd/wag, compared with homozygous ACI/ACI hybrids, reflecting the effect of WAG alleles at the locus in limiting the duration of a train of SWDs. Both these results and the negative product–moment correlations between type I and II SWDs in the cross support our hypothesis that increased type II SWD activity is accompanied by a reduced occurrence of type I SWD or limits the duration of an ongoing type I SWD. The fact that apomorphine has opposite effects on type I and type II SWDs is in line with this hypothesis.
The characterisation of the T1swd/wag and T2swd/wag with respect to their specific effects on type I and type II SWDs provides important information for the selection of candidate genes at the loci. The negative correlations between type I and II SWDs and their opposite reaction to a single drug may connect both loci to the DA system. Both rat genome annotation (http://www.ensembl.org/Rattus_norvegicus/) and comparative mapping data between rat, mouse, and human genomes (25–27, http://www.well.ox.ac.uk/rat_mapping_resources) generate crucial information on candidate genes for T1swd/wag and T2swd/wag and allow comparisons between genetic-mapping results of epilepsy-related phenotypes across species. The QTL regions contain genes encoding ion channels that are potential candidates for neurologic disorders, including Slc4a3, an anion exchanger (43), Slc9a1, a sodium/hydrogen exchanger, and Kcnab2, a potassium voltage-gated channel of the shaker-related subfamily (44). Interestingly, in the slow-wave epilepsy (swe) mouse, a mutation in Slc9a1 causes an epilepsy-like seizure phenotype (45). In addition, the gene encoding the tissue nonspecific alkaline phosphatase (TNAP), which maps to the locus T2swd/wag, plays a role in seizures in homozygous mutant mice deficient for the gene (46). The QTL T1swd/wag maps to a region conserved with the mouse QTL Szs5 controlling seizure responses to pentylenetetrazol in a DBA/2JxC57BL/6J cross (47). This rat QTL also may share homology relations with human chromosome 2q36 marginally linked to epilepsy in CAE patients (48).
Our QTL mapping data demonstrate the polygenic control of SWD phenotypes in a WAG/Rij × ACI strain combination and have significant implications for the study of mechanisms underlying the independent genetic control of type I and type II SWDs, as well as the differential effect of gene variants at the loci on the average duration and number of SWDs. The production and characterization of congenic lines for the QTLs, combined with microarray-based gene-transcription profiling and computational analysis of the rat genome assembly, is the main strategy for fine QTL mapping and isolating candidate genes. Overall, results from the genetic dissection of SWD phenotypes in the WAG/Rij rat should lead to fundamental new insights into endogenous mechanisms that trigger spontaneously occurring SWDs.
Acknowledgment: We acknowledge the biotechnical assistance of Jean-Paul Dibbets, Hans Krijnen, and Elly Willems van Bree. Dr. Julia Klioueva assisted in long-term EEG recordings. Work in Oxford was supported by the Wellcome Trust. D.G. holds a Wellcome Trust senior fellowship in basic biomedical science (057733). S.P.W. is a recipient of a Wellcome Prize Studentship in Bioinformatics and Statistical Genetics.
- 1The epidemiology and prognosis of typical absence seizures. In: DuncanJS, PanayiotopoulosCP, eds. Typical absences and related epileptic syndromes. London : Churchill Communications, 1995: 135–44..
- 23The rat brain in stereotaxoc coordinates. London : Academic Press, 1997., .
- 27A map of the rat genome containing 5,203 markers: 4,700 microsatellites and 605 genes in a rat, mouse and human comparative map. Nat Genet 1999;2: 27–36., , , et al.
- 28JoinMap version 2.0: Software for the calculation of genetic linkage maps: centre for Plant Breeding and Reproduction Research. Wageningen, The Netherlands : 1995., .