November 2008
Volume 49, Issue 11
Free
Clinical and Epidemiologic Research  |   November 2008
A Canine Model of Inherited Myopia: Familial Aggregation of Refractive Error in Labrador Retrievers
Author Affiliations
  • Joanna Black
    From the Departments of Optometry and Vision Science and
  • Sharon R. Browning
    Statistics, University of Auckland, Auckland, New Zealand.
  • Andrew V. Collins
    From the Departments of Optometry and Vision Science and
  • John R. Phillips
    From the Departments of Optometry and Vision Science and
Investigative Ophthalmology & Visual Science November 2008, Vol.49, 4784-4789. doi:10.1167/iovs.08-1828
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to Subscribers Only
      Sign In or Create an Account ×
    • Get Citation

      Joanna Black, Sharon R. Browning, Andrew V. Collins, John R. Phillips; A Canine Model of Inherited Myopia: Familial Aggregation of Refractive Error in Labrador Retrievers. Invest. Ophthalmol. Vis. Sci. 2008;49(11):4784-4789. doi: 10.1167/iovs.08-1828.

      Download citation file:


      © 2016 Association for Research in Vision and Ophthalmology.

      ×
  • Supplements
Abstract

purpose. To determine whether the distribution of naturally occurring myopia in Labrador Retrievers has a genetic component.

methods. Pedigree records of a large canine family were analyzed. Pure Labrador Retrievers, 1 to 8 years of age, free of ocular disease, and available for testing were studied. Refractive error was measured by cycloplegic retinoscopy in both eyes. The family included mating loops, and so an expectation maximization (EM) algorithm (multivar program, MORGAN software; University of Washington, Seattle) was used to calculate log likelihoods of refractive error with environmental and additive genetic models. The fixed effects of coat color, sex, and litter size were also tested.

results. In our sample of 116 dogs from this one family, the average spherical equivalent refraction (SER) was −0.41 D (range, −5.38 to +1.65 D, mean of both eyes, n = 116): 31% were myopic (SER ≤ −0.50 D), 60% were emmetropic (SER = −0.49 to +0.99 D), and 9% were hyperopic (SER ≥ +1.00 D). The significance of fixed and genetic effects was tested by comparing the full model (including genetic and all fixed effects) to models with one effect removed. Litter size and additive genetic effects were significant (P = 0.0013 and P = 0.000093, respectively), whereas sex and coat color were not. The overall variance in SER was accounted for approximately equally by additive genetic variance and residual/environmental variance. Narrow sense heritability of SER was 0.506.

conclusions. The distribution of refractive error within this family of Labrador Retrievers had a significant genetic component, but was also influenced by other factors (litter size, and undefined residual/environmental effects). The dog represents a unique model for the study of naturally occurring, heritable, high-prevalence, low-degree myopia.

A variety of animals, including the chick, 1 guinea pig, 2 tree shrew, 3 marmoset, 4 and monkey 5 have been used as animal models of human myopia. In these models, the eyes are not naturally myopic: rather, myopia is induced in developing eyes by manipulation of the visual environment. Myopia is typically induced by imposing retinal defocus (with negative lens wear) or by visual form deprivation, both of which cause axial elongation of the vitreous chamber and myopia in all eyes. Furthermore, when the inducing stimulus is removed, the myopia recovers over time. 1 The link between these animal models and myopia commonly found within the human population is questionable because human myopia, including high myopia and the much more common juvenile onset myopia, both appear to have a significant genetic component. Children with myopic parents have a higher than normal risk of myopia, 6 7 8 and high between-sibling correlations for refractive error and progression of myopia have been reported. 9 10 Also, twin studies show a higher level of concordance of common myopia in monozygotic compared to dizygotic twins. 6 7 The clear familial nature of high myopia 8 (refractive error ≤ −6.00 D) has led to many investigations of chromosomal regions linked to high myopia in humans, which have identified possible candidate loci at Xq28 (MYP1), 11 18p11 (MYP2), 12 12q21 to 23 (MYP3), 13 7q36 (MYP4), 14 17q21-22 (MYP5), 15 4q22-27 (MYP11), 16 2q37 (MYP12), 17 and Xq23-25 (MYP13). 18 Recent studies have also investigated the genetics of common myopia, which is more prevalent in the population. These studies have revealed possible candidate regions at 22q12 (MYP6), 19 11p13 (MYP7), 20 3q26 (MYP8), 20 4q12 (MYP9), 20 8p23 (MYP10), 20 and 1p36 (MYP14). 21 Families from a range of ethnicities and geographic backgrounds have been investigated in studies of high myopia and common myopia, and the results have not always been the same. 
One animal species in which myopia occurs naturally is the domestic dog. Although the prevalence of myopia in dogs is breed dependent, approximately 8% to 15% of Labrador Retrievers are reported to have myopia ranging from −0.50 to −5.00 D. 22 23 The fact that the dog is the only animal species identified as having significant levels of naturally occurring myopia is interesting, but what is not clear is whether that myopia has a heritable component. If it does, then the dog may serve as a powerful animal model of human myopia. Moreover, the dog would have several advantages over humans for the study of the genetics of myopia. The use of dogs could minimize the confounding environmental effects of near work, education levels, and mode of upbringing typically identified as risk factors in humans. 24 Pedigree dogs also provide well-documented multigeneration ancestry that is often unknown or uncertain in human families. Furthermore, dogs mature relatively rapidly and have large litters and a high reproduction rate, so that many individuals from different generations can be studied relatively quickly. 
The purpose of this study was to evaluate the inheritance of myopia in the dog as a prelude to possible genotyping. We have studied the Labrador Retriever because it is the only dog breed in which myopia has been shown to be based on vitreous chamber elongation, 22 as is the case in human myopia. Several studies have used segregation analysis methods to investigate the mode of inheritance of other canine diseases, such as deafness 25 and hip dysplasia. 26 However, the Statistical Analysis for Genetic Epidemiology (SAGE; 1997) software package that they used (Case Western University, Cleveland, OH) was not designed to cope with consanguinity and multiple matings within the sample population. We have chosen to use an expectation maximization (EM) algorithm that can fit an additive variance components genetic model to large pedigrees with multiple mates and consanguinity, 27 because our sample population contained many such loops. 
Methods
All procedures within this study complied with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Animal Ethics Committee of The University of Auckland. 
Study Population
Dogs were identified from the records of a colony specializing in breeding service dogs, situated in Auckland, New Zealand. Of the 1342 dogs in the colony’s records, 76% were Labrador Retrievers, registered with The New Zealand Kennel Club Purebred Dog Registry. Breeding and veterinary records, including regular ophthalmic records from 6 weeks of age onward, were available for all animals. Because all originated from the same colony, all dogs in the study had experienced a similar visual environment to the extent that, as puppies, they remained indoors with regular time outdoors from birth until 6 weeks of age. They were then separated from their littermates to live in a domestic environment. 
A large family of Labrador Retrievers originating from a stud dog known to have myopia (−2.00 D) and having 407 descendents in 64 litters was selected for study. Several exclusion criteria were used in selecting the dogs from this family for further analysis. First, animals outside the age range of 1 to 8 years were excluded. This age range is developmentally and physiologically equivalent to a human age range of 18 to 56 years for dogs the size of Labrador Retrievers. 28 The upper age limit (8 years) was selected to reduce the likelihood of including dogs with lenticular changes that may induce a myopic shift in refraction. 23 Animals under 1 year of age were excluded on the grounds that while the rate of refractive development in dogs is unknown, dogs reach sexual maturity between 6 and 12 months of age. 29 Thus, by 1 year of age (equivalent to 18 years of age in humans), their refractive status would be expected to be stable. Dogs were excluded from the study, or were not able to be tested for a variety of reasons which are detailed in Table 1 . All dogs that were phenotyped were included in the study. Selection bias was minimized, as there were no behavioral or other prior indicators of the refractive error of any of the dogs. In total, 116 dogs were included in the study, as they were registered as pure Labrador Retrievers within the specified age range, were members of the same extended family, were free of ocular disease, were traceable and were available for testing. Pedigree drawing software (CraneFoot, ver. 3.2.2; Helsinki University of Technology) was used to map out phenotyped individuals and mating loops within the sample pedigree. 
Refractive Error
The refractive error of each eye of 116 dogs was measured under cycloplegia with streak retinoscopy, independently by two retinoscopists. Dogs tested for refractive error also had a buccal swab taken for later genetic analysis. Cycloplegia was induced with 1% tropicamide (Mydriacyl; Alcon Laboratories, Fort Worth, TX) instilled 30 minutes before the retinoscopy measures. 30 The results of each retinoscopist were recorded as spherical equivalent refraction (SER), the mean of the refractions measured in the vertical and horizontal power meridians. The results of the two retinoscopists were then averaged for each eye. For the purpose of analysis and modeling, the SER assigned to each dog was the mean SER of both eyes. For the purpose of computing myopia prevalence, we defined myopia as an SER (mean of both eyes) of ≤ −0.50 D, emmetropia as −0.49 to +0.99 D, and hyperopia as ≥ +1.00 D, as these limits have typically been used in other human 31 and canine 22 studies. We emphasize that these definitions of myopia, emmetropia, and hyperopia had no bearing on the variance component modeling of quantitative traits. For modeling purposes, the numerical value of SER assigned to each dog was simply the mean SER of both eyes. 
Data Analysis
MORGAN software version 2.8.2 (University of Washington, Seattle, WA) was used to analyze the data because it allowed analysis of large pedigrees with multiple mates and consanguinity, which characterized our sample. Our analysis accounted for all known relationships between the phenotyped dogs and their shared ancestors. 
The pedcheck program of MORGAN was used to check for relationship errors within the pedigree. The multivar program of MORGAN was used to obtain maximum-likelihood estimates for additive polygenic models of quantitative traits by the expectation-maximization (EM) algorithm; this program did not allow other (e.g., dominant) models to be investigated. We analyzed the inheritance of refractive error as a quantitative trait, with three fixed effects: sex, litter size, and coat color. We included sex, as there is some suggestion of a sex effect in human studies. 32 Although there is no previous literature to suggest that either coat color or litter size would be associated with myopia, both are phenotypic descriptors that do not alter with age. Both were available for all dogs and were included because of the chance that they might be relevant. The MORGAN package was also used to calculate inbreeding coefficients between individuals of interest, by using the kin program. The inbreeding coefficient (F) is a measure of the probability of homozygosity by descent (i.e., the chance of an animal’s inheriting the same ancestral gene from both parents because they are related). Narrow sense heritability (the additive genetic variance as a proportion of total variance) was calculated using variances computed by multivar for the best fitting model only. 
The additive model used by multivar 27 is y = μ + z + e, where y is the quantitative trait, μ is a vector of fixed effects, z is a vector of additive genetic effects, and e is a vector of residual/environmental effects. The distributions of e and z are assumed to be normal. The elements of e are assumed to be independent with common variance. The elements of z have correlations proportional to the coefficients of kinship. The multivar program fits the model using the EM algorithm, which is an iterative approach to finding the parameter values that maximize the model likelihood. The stability of parameter values was typically achieved by 50 iterations, although we always used 200 iterations, which is the default for multivar
A histogram of refractive error distribution within the sample showed that it was left skewed. Approximate normality was obtained by the transformation √(1.65 − SER). This transformation was chosen to obtain positive right-skewed values first (the largest untransformed SER value was +1.65 D), and then to reduce the skewness through use of the square-root transformation. The transformed values were used as the quantitative trait in subsequent analysis with the multivar program. We also ran the analysis using a normal deviates transform 33 of the refractive error distribution. Litter size ranged between 1 and 13 puppies and was coded as small (litter size between 1 and 7) or large (between 8 and 13). 
To test for statistical significance of the fixed and genetic effects, we compared nested models. For example, we compared the full model with genetic effects, sex, coat color, and litter size to submodels, in which each one of these effects was removed in turn. The null hypothesis was that the submodel was as good as the full model at explaining the variability of the trait. Twice the difference in log likelihoods from the two models was compared to a χ2 distribution with degrees of freedom equal to the difference in the number of estimated parameters between the two models. As each of our fixed effects had two possible states (apart from missing values) the degrees of freedom was equal to one in each case. If the resulting probability was significant (P < 0.05), the null hypothesis was rejected, and we concluded that the variable that was removed from the full model to create the submodel was statistically significant in helping to explain the variability of the trait. 
Results
Pedigree
A total of 116 dogs in 37 related litters were tested for refractive error, as they were registered as pure Labrador Retrievers within the age range of 1 to 8 years, had no known ocular disease, and were available for phenotypic analysis (Table 1) . The MORGAN multivar program required all animals in the analysis to have both parents included, even if their refractive error had not been measured. In these cases, the SER was entered as unknown, but the fixed effects (their sex, coat color, and litter size) were known and were entered. In the one case where only one parent was known, a surrogate parent was entered with no fixed-effect data. As a result, a total of 134 dogs were included in the analysis, of which 116 had known refractions. Figure 1shows the relationships between the 134 dogs, including the mating loops. The sample included 62 females and 72 males, with 92 dogs having yellow, 37 black and 5 unknown coat color. Although Figure 1shows dogs from 37 litters, it does not show all dogs in each litter. Refractive error was obtained from an average of 42% of dogs from each litter, and only those whose refractive errors were determined are shown in Figure 1 . When all dogs in each of the 37 litters were included in the computation of litter sizes, the average litter size was 7 (range, 1–13). The mean inbreeding coefficient (F) in the sample tested was 2.05% (range, 12.5%– 0.0%). In the total family (407 dogs), mean F = 2.00% (range, 12.5%–0.0%). 
Refractive Error Distribution
Refractive error was measured by two independent retinoscopists using cycloplegic retinoscopy. The 95% limits of agreement for the results of the two retinoscopists was 1.10 D. After the results obtained by the two retinoscopists for each eye were averaged, the correlation of SER between eyes was R = 0.963. The average SER for the 116 dogs that were tested was −0.41 D (range, +1.65 to −5.38 D). The average difference in SER between the right and left eyes was 0.13 D, with anisometropia of ≥1.00 D present in seven dogs (range, 1 to 2.05 D). The levels of astigmatism within the sample were generally low. The average astigmatism (difference between vertical and horizontal meridional powers) for all 232 eyes was 0.028 D (range, 0.00 to 2.60 D). Nine eyes (from six animals) had astigmatism > 0.50 D and of those, three eyes from two animals had astigmatism > 1.00 D. 
Figure 2shows the distribution of refractive error in our sample (n = 116). The distribution appeared bimodal, with most dogs having refractive errors clustered around low hyperopia (+0.50 D), whereas dogs with myopia appeared to form a group centered around −2.50 D. Within this sample, 31% were myopic (SER ≤ −0.50 D), 60% were emmetropic (SER = −0.49 to +0.99 D), and 9% were hyperopic (SER ≥ +1.00 D). Figure 3shows refractive errors of the 116 dogs plotted versus age. In this sample, ranging in age from 1 to 8 years, there was no correlation of SER with age (Pearson R = 0.0047, P = 0.52). 
Familial Aggregation
The statistical significance of fixed effects (sex, coat color, and litter size) and genetic effects was tested by comparing the full model (with genetic and all fixed effects) to models with one of these effects removed (Table 2) . Litter size and additive genetic effects were significant (P = 0.0013 and P = 0.000093, respectively), whereas sex and coat color were not significant (P = 0.20 and P = 0.15, respectively). Similar results (in terms of statistical significance of each effect) were obtained by comparing a null model (with residual/environmental effects only) to models with only one of the fixed or genetic effects added. Analysis using the normal deviates transform confirmed the significance of litter size and additive genetic effects. 
The final model, with litter size and genetic effects, had the form:  
\[f(y){=}{\mu}{+}{\alpha}_{1}1{\{}\mathrm{small}{\}}{+}{\alpha}_{2}1{\{}\mathrm{large}{\}}{+}z{+}e\]
where y is the SER, f is the transformation to improve normality f(y) = √(1.65 − y); μ is the overall mean; 1{small} is 1 if the individual comes from a small litter and 0 otherwise; while 1{large} is 1 if the individual comes from a large litter and 0 otherwise (each individual comes from either a small or a large litter, so 1{small} and 1{large} always sum to 1); α1 and α2 give the effects of small or large litter size; z is the additive genetic effect, which has a normal distribution with mean 0 and variance σg 2 (with correlations between individuals being described by the matrix of kinship coefficients); and e is the residual/environmental effect, which has a normal distribution with mean 0 and variance σe 2. The fitted values for these parameters were: μ = 1.218, α1 = 0.112, α2 = −0.062, σg 2 = 0.128, and σe 2 = 0.118. There were 38 dogs from small litters and 68 from large litters, and the variance for litter size was 0.007. Thus, total variance was 0.128 + 0.118 + 0.007 = 0.253, and narrow-sense heritability was 0.128/0.253 = 0.506 (computed from the best-fitting values). 
The overall variance after accounting for litter size was split approximately equally between additive genetic variance and residual/environmental variance. Thus, genetic and undefined residual/environmental factors played approximately equal roles in determining refractive error in this family. Litter size made a much smaller contribution to refractive error. Dogs from small litters had an average transformed trait value of 1.218 + 0.112 = 1.330, which corresponds to SER = 1.65 − 1.332 = −0.119 D. On the other hand, dogs from large litters had an average transformed trait value of 1.218 − 0.062 = 1.156, which corresponds to SER = 1.65 − 1.1562 = 0.314 D. Thus, the model indicated that dogs from large litters had refractive errors that were, on average, 0.43 D more hyperopic than dogs from small litters. 
Discussion
The results of our study indicate that myopia in Labrador Retrievers has a significant genetic component. To our knowledge, this is the first study to demonstrate a genetic component responsible for naturally occurring myopia in any species other than humans. A large number of animals of several species including monkeys have been examined during studies of refractive development. 1 2 3 4 5 Variability in response (e.g., to lens wear) among individual animals has sometimes been attributed to differences in genetic susceptibility to retinal defocus. 3 However, there have been no consistent reports of naturally occurring myopia in any animals other than dogs. 
In the dogs in our study, the prevalence of myopia (SER ≤ −0.50 D) was 31%, which is much greater than that reported in Labrador Retrievers by Mutti et al., 22 who found that 8% of their sample of 75 dogs were myopic in both eyes by at least −0.50 D. However, this difference most likely reflects the fact that our sample was selected from a family with a known myopic progenitor. This is a limitation of our study that restricts generalization of our myopia prevalence data to within our sample. Of interest, the ranges of refractive error found for dogs in this study (+1.65 to −5.38 D) and that of Mutti et al. (+3.50 to −5.00 D) 22 were similar, particularly for the limit of myopia. Therefore, the data from both studies on prevalence and degree of myopia suggest that myopia in dogs more closely resembles common myopia rather than high myopia in humans. Human high myopia typically has a prevalence of 1% to 2% and a degree of −6.00 D or more minus, whereas common human myopia has a variable prevalence (10%–75%) with a degree less minus than −6.00 D. 8 Thus, dogs appear to provide a unique animal model for examining the genetics of common (i.e., high-prevalence, low-degree) myopia. 
Dogs provide an interesting perspective on development of myopia. In the Labrador Retriever family that we studied, additive genetic effects accounted for about half of the variance in SER, a level similar to that reported in human family studies, 32 34 although much less than the high levels reported in twin studies in which additive genetic effects account for more than 80% of variance in refraction. 35 36  
Many service dogs have a domestic environment comparable to that of humans and also enjoy a good-quality diet and a high level of medical care. However, dogs do not experience many of the environmental influences that have been implicated in the development of myopia in humans (i.e., prolonged near work like reading). Dogs are not subject to near work, which may suggest that myopia in susceptible dogs has developed as a response to some environmental influence other than near work, for example an indoors environment. However, valid comparison between dogs and humans is difficult. The amplitude of accommodation in the dog is reportedly 2 to 3 D, 37 whereas that of a child is much greater (e.g., for a 10-year-old child it is typically approximately 12 D 38 ). Thus, in terms of accommodative demand relative to available amplitude (and possibly consequent lag of accommodation 39 ) a near stimulus (e.g., demanding half the available amplitude of accommodation) would correspond to a stimulus at approximately 17 cm distance for a 10-year-old child but approximately 1.0 m distance for a dog. 
An interesting finding from this study was that dogs from small litters were likely to be more myopic than dogs from large litters. The origin of this effect remains unclear. It may suggest some link between the genetic determinants of litter size and those of refractive error. Alternatively, it may result from an environmental effect. Although all dogs in the study were raised in the same indoors environment from birth to 6 weeks of age, dogs from small litters may have developed more rapidly under these early myopiagenic conditions than dogs from large litters, which may have developed later. Another possibility is that the myopiagenic environments were the same for both litter types, but greater levels of physical activity were required to live and feed in large litters than in small ones. It has been reported recently that increased levels of activity may have a protective effect on progression of myopia in humans. 31 40  
Our statistical analysis was complicated by the processing of consanguineous relationships and multiple pairings. The multivar program used the trait information for all members of the pedigree rather than having to split the family into smaller nuclear groups, as has been done in some previous studies. 25 26 However, our analysis method did not allow study of the mode of transmission, although the finding that sex was not significant implies that myopia in dogs has an autosomal inheritance pattern. 
In conclusion, dogs may provide a uniquely useful animal model for examining the genetics of common myopia, as they do for many other diseases. 41 42 This finding is particularly relevant after the completion and public availability of the dog genome. 42 Moreover, several microsatellite marker screening sets 43 44 45 46 have also been completed for use in canine genome scans in linkage studies, and linkage studies have successfully identified causative genes for canine eye disease such as Collie eye anomaly. 47 Furthermore, studies of environmental influences on puppy refractive development may provide insight into the subtle environmental factors responsible for the emergence of common myopia in genetically susceptible young humans. 
 
Table 1.
 
Status of Dogs in the Family of Labrador Retrievers Studied, at the Time of Analysis
Table 1.
 
Status of Dogs in the Family of Labrador Retrievers Studied, at the Time of Analysis
Total refracted 116
Outside age range (<1 or >8 y) 131
In New Zealand but not available for testing 71
Outside New Zealand 40
Ocular disease 24
Cross-breed 14
Deceased 11
Total family 407
Figure 1.
 
Pedigree diagram showing the relationships between dogs in the 37 litters studied and the refractive error status of the dogs. Myopia is SER ≤ −0.50 D; emmetropia is SER between −0.49 and +0.99 D; and hyperopia is SER ≥ +1.00 D. Within this sample, 31% were myopic, 60% were emmetropic, and 9% were hyperopic. The common ancestor (proband) is represented multiple times (eight times) by a small black pointer. Curved lines show repeat instances of the same dog(s) within the pedigree.
Figure 1.
 
Pedigree diagram showing the relationships between dogs in the 37 litters studied and the refractive error status of the dogs. Myopia is SER ≤ −0.50 D; emmetropia is SER between −0.49 and +0.99 D; and hyperopia is SER ≥ +1.00 D. Within this sample, 31% were myopic, 60% were emmetropic, and 9% were hyperopic. The common ancestor (proband) is represented multiple times (eight times) by a small black pointer. Curved lines show repeat instances of the same dog(s) within the pedigree.
Figure 2.
 
Distribution of refractive error (mean SER of both eyes) within the sample of 116 dogs who were all descendents of one stud dog with myopia of −2.00 D. Histogram bin width, 0.50 D. Mean refractive error of this group = −0.41 D (range, −5.38 to +1.65 D).The bimodal nature of this distribution suggests that dogs with myopia tended to form a distinct group, with the major group having a near-normal distribution about low hyperopia.
Figure 2.
 
Distribution of refractive error (mean SER of both eyes) within the sample of 116 dogs who were all descendents of one stud dog with myopia of −2.00 D. Histogram bin width, 0.50 D. Mean refractive error of this group = −0.41 D (range, −5.38 to +1.65 D).The bimodal nature of this distribution suggests that dogs with myopia tended to form a distinct group, with the major group having a near-normal distribution about low hyperopia.
Figure 3.
 
Plot of refractive error (mean SER of both eyes) versus age for the 116 dogs in the sample. All dogs had ages within the range 1 to 8 years. Shown is the best fitting (least squares) linear fit to the data (SER = 0.003(age) + 0.42). There was no significant relationship between refractive error and age for the dogs in this sample (Pearson R = 0.0047, P = 0.52).
Figure 3.
 
Plot of refractive error (mean SER of both eyes) versus age for the 116 dogs in the sample. All dogs had ages within the range 1 to 8 years. Shown is the best fitting (least squares) linear fit to the data (SER = 0.003(age) + 0.42). There was no significant relationship between refractive error and age for the dogs in this sample (Pearson R = 0.0047, P = 0.52).
Table 2.
 
Models Used to Test for Statistical Significance of Variables
Table 2.
 
Models Used to Test for Statistical Significance of Variables
Model Comment Log Likelihood χ2 P
G + S + C + L + E Full model −66.50 N/A N/A
S + C + L + E Test for additive genetic effect −74.14 15.28 9.3× 10−5 *
G + C + L + E Testing for effect of sex −67.31 1.62 0.20
G + S + L + E Testing for effect of coat color −67.56 2.12 0.15
G + S + C + E Testing for effect of litter size −71.64 10.28 0.0013*
The authors thank the staff at the breeding colony, in particular Nicola Cadogan, and also Scott Bruce, Joe Walker, Janice Lloyd, Ian Cox, Denise Ireland, Kim Malcolm, and Helen Grant for assistance; Elizabeth Thompson for help related to the multivar program; Nicola Anstice and Simon Backhouse for help with refractive error testing; and Andrew Shelling for advice. 
WallmanJ, AdamsJI. Developmental aspects of experimental myopia in chicks: susceptibility, recovery and relation to emmetropization. Vision Res. 1987;27:1139–1163. [CrossRef] [PubMed]
HowlettMH, McFaddenSA. Form-deprivation myopia in the guinea pig (Cavia porcellus). Vision Res. 2006;46:267–283. [CrossRef] [PubMed]
NortonTT, SiegwartJT, Jr, AmedoAO. Effectiveness of hyperopic defocus, minimal defocus, or myopic defocus in competition with a myopiagenic stimulus in tree shrew eyes. Invest Ophthalmol Vis Sci. 2006;47:4687–4699. [CrossRef] [PubMed]
TroiloD, NicklaDL, WildsoetCF. Form deprivation myopia in mature common marmosets (Callithrix jacchus). Invest Ophthalmol Vis Sci. 2000;41:2043–2049. [PubMed]
SmithEL, 3rd, RamamirthamR, Qiao-GriderY, et al. Effects of foveal ablation on emmetropization and form-deprivation myopia. Invest Ophthalmol Vis Sci. 2007;48:3914–3922. [CrossRef] [PubMed]
LyhneN, SjolieAK, KyvikKO, GreenA. The importance of genes and environment for ocular refraction and its determiners: a population based study among 20-45 year old twins. Br J Ophthalmol. 2001;85:1470–1476. [CrossRef] [PubMed]
DiraniM, ChamberlainM, GaroufalisP, ChenC, GuymerRH, BairdPN. Refractive errors in twin studies. Twin Res Hum Genet. 2006;9:566–572. [CrossRef] [PubMed]
GuggenheimJA, KirovG, HodsonSA. The heritability of high myopia: a reanalysis of Goldschmidt’s data. J Med Genet. 2000;37:227–231. [CrossRef] [PubMed]
SawSM, NietoFJ, KatzJ, ScheinOD, LevyB, ChewSJ. Familial clustering and myopia progression in Singapore school children. Ophthalmic Epidemiol. 2001;8:227–236. [CrossRef] [PubMed]
GuggenheimJA, Pong-WongR, HaleyCS, GazzardG, SawSM. Correlations in refractive errors between siblings in the Singapore Cohort Study of Risk factors for myopia. Br J Ophthalmol. 2007;91:781–784. [CrossRef] [PubMed]
HaimM, FledeliusHC, Skarsholm . X-linked myopia in Danish family. Acta Ophthalmol (Copenh). 1988;66:450–456. [PubMed]
YoungTL, RonanSM, DrahozalLA, et al. Evidence that a locus for familial high myopia maps to chromosome 18p. Am J Hum Genet. 1998;63:109–119. [CrossRef] [PubMed]
YoungTL, RonanSM, AlvearAB, et al. A second locus for familial high myopia maps to chromosome 12q. Am J Hum Genet. 1998;63:1419–1424. [CrossRef] [PubMed]
NaiglinL, GazagneC, DallongevilleF, et al. A genome wide scan for familial high myopia suggests a novel locus on chromosome 7q36. J Med Genet. 2002;39:118–124. [CrossRef] [PubMed]
PaluruP, RonanSM, HeonE, et al. New locus for autosomal dominant high myopia maps to the long arm of chromosome 17. Invest Ophthalmol Vis Sci. 2003;44:1830–1836. [CrossRef] [PubMed]
ZhangQ, GuoX, XiaoX, JiaX, LiS, HejtmancikJF. A new locus for autosomal dominant high myopia maps to 4q22–q27 between D4S1578 and D4S1612. Mol Vis. 2005;11:554–560. [PubMed]
PaluruPC, NallasamyS, DevotoM, RappaportEF, YoungTL. Identification of a novel locus on 2q for autosomal dominant high-grade myopia. Invest Ophthalmol Vis Sci. 2005;46:2300–2307. [CrossRef] [PubMed]
ZhangQ, GuoX, XiaoX, JiaX, LiS, HejtmancikJF. Novel locus for X linked recessive high myopia maps to Xq23–q25 but outside MYP1. J Med Genet. 2006;43:e20. [PubMed]
StambolianD, IbayG, ReiderL, et al. Genomewide linkage scan for myopia susceptibility loci among Ashkenazi Jewish families shows evidence of linkage on chromosome 22q12. Am J Hum Genet. 2004;75:448–459. [CrossRef] [PubMed]
HammondCJ, AndrewT, MakYT, SpectorTD. A susceptibility locus for myopia in the normal population is linked to the PAX6 gene region on chromosome 11: a genomewide scan of dizygotic twins. Am J Hum Genet. 2004;75:294–304. [CrossRef] [PubMed]
WojciechowskiR, MoyC, CinerE, et al. Genomewide scan in Ashkenazi Jewish families demonstrates evidence of linkage of ocular refraction to a QTL on chromosome 1p36. Hum Genet. 2006;119:389–399. [CrossRef] [PubMed]
MuttiDO, ZadnikK, MurphyCJ. Naturally occurring vitreous chamber-based myopia in the Labrador retriever. Invest Ophthalmol Vis Sci. 1999;40:1577–1584. [PubMed]
MurphyCJ, ZadnikK, MannisMJ. Myopia and refractive error in dogs. Invest Ophthalmol Vis Sci. 1992;33:2459–2463. [PubMed]
MuttiDO, MitchellGL, MoeschbergerML, JonesLA, ZadnikK. Parental myopia, near work, school achievement, and children’s refractive error. Invest Ophthalmol Vis Sci. 2002;43:3633–3640. [PubMed]
CargillEJ, FamulaTR, StrainGM, MurphyKE. Heritability and segregation analysis of deafness in U.S. Dalmatians. Genetics. 2004;166:1385–1393. [CrossRef] [PubMed]
JanuttaV, HamannH, DistlO. Complex segregation analysis of canine hip dysplasia in German shepherd dogs. J Hered. 2006;97:13–20. [CrossRef] [PubMed]
ThompsonEA, ShawRG. Pedigree analysis for quantitative traits: variance components without matrix inversion. Biometrics. 1990;46:399–413. [CrossRef] [PubMed]
WatersDJ, PatronekGJ, BostwickDG, GlickmanLT. Comparing the age at prostate cancer diagnosis in humans and dogs. J Natl Cancer Institute. 1996;88:1686–1687. [CrossRef]
KutzlerMA. Induction and synchronization of estrus in dogs. Theriogenology. 2005;64:766–775. [CrossRef] [PubMed]
MuttiDO, ZadnikK, EgashiraS, KishL, TwelkerJD, AdamsAJ. The effect of cycloplegia on measurement of the ocular components. Invest Ophthalmol Vis Sci. 1994;35:515–527. [PubMed]
JacobsenN, JensenH, GoldschmidtE. Does the level of physical activity in university students influence development and progression of myopia?—a 2-year prospective cohort study. Invest Ophthalmol Vis Sci. 2008;49:1322–1327. [CrossRef] [PubMed]
ChenCY, ScurrahKJ, StankovichJ, et al. Heritability and shared environment estimates for myopia and associated ocular biometric traits: the Genes in Myopia (GEM) family study. Hum Genet. 2007;121:511–520. [CrossRef] [PubMed]
LegendrePL. Numerical Ecology. 1998; 2nd ed.Elsevier Amsterdam, The Netherlands.
BiinoG, PalmasMA, CoronaC, et al. Ocular refraction: heritability and genome-wide search for eye morphometry traits in an isolated Sardinian population. Hum Genet. 2005;116:152–159. [CrossRef] [PubMed]
HammondCJ, SniederH, GilbertCE, SpectorTD. Genes and environment in refractive error: the twin eye study. Invest Ophthalmol Vis Sci. 2001;42:1232–1236. [PubMed]
DiraniM, ChamberlainM, ShekarSN, et al. Heritability of refractive error and ocular biometrics: the Genes in Myopia (GEM) twin study. Invest Ophthalmol Vis Sci. 2006;47:4756–4761. [CrossRef] [PubMed]
MillerPE, MurphyCJ. Vision in dogs. J Am Vet Med Assoc. 1995;207:1623–1634. [PubMed]
SternerB, GellerstedtM, SjostromA. The amplitude of accommodation in 6–10-year-old children—not as good as expected!. Ophthalmic Physiol Opt. 2004;24:246–251. [CrossRef] [PubMed]
GwiazdaJ, ThornF, BauerJ, HeldR. Myopic children show insufficient accommodative response to blur. Invest Ophthalmol Vis Sci. 1993;34:690–694. [PubMed]
RoseKA, MorganIG, IpJ, et al. Outdoor activity reduces the prevalence of myopia in children. Ophthalmology. 2008;115:1279–1285. [CrossRef] [PubMed]
SarganDR. IDID: inherited diseases in dogs: web-based information for canine inherited disease genetics. Mamm Genome. 2004;15:503–506. [CrossRef] [PubMed]
OstranderEA, WayneRK. The canine genome. Genome Res. 2005;15:1706–1716. [CrossRef] [PubMed]
RichmanM, MellershCS, AndreC, GalibertF, OstranderEA. Characterization of a minimal screening set of 172 microsatellite markers for genome-wide screens of the canine genome. J Biochem Biophys Methods. 2001;47:137–149. [CrossRef] [PubMed]
SarganDR, Aguirre-HernandezJ, GalibertF, OstranderEA. An extended microsatellite set for linkage mapping in the domestic dog. J Hered. 2007;98:221–231. [CrossRef] [PubMed]
GuyonR, LorentzenTD, HitteC, et al. A 1-Mb resolution radiation hybrid map of the canine genome. Proc Natl Acad Sci U S A. 2003;100:5296–5301. [CrossRef] [PubMed]
ClarkLA, TsaiKL, SteinerJM, et al. Chromosome-specific microsatellite multiplex sets for linkage studies in the domestic dog. Genomics. 2004;84:550–554. [CrossRef] [PubMed]
LoweJK, KukekovaAV, KirknessEF, et al. Linkage mapping of the primary disease locus for collie eye anomaly. Genomics. 2003;82:86–95. [CrossRef] [PubMed]
Figure 1.
 
Pedigree diagram showing the relationships between dogs in the 37 litters studied and the refractive error status of the dogs. Myopia is SER ≤ −0.50 D; emmetropia is SER between −0.49 and +0.99 D; and hyperopia is SER ≥ +1.00 D. Within this sample, 31% were myopic, 60% were emmetropic, and 9% were hyperopic. The common ancestor (proband) is represented multiple times (eight times) by a small black pointer. Curved lines show repeat instances of the same dog(s) within the pedigree.
Figure 1.
 
Pedigree diagram showing the relationships between dogs in the 37 litters studied and the refractive error status of the dogs. Myopia is SER ≤ −0.50 D; emmetropia is SER between −0.49 and +0.99 D; and hyperopia is SER ≥ +1.00 D. Within this sample, 31% were myopic, 60% were emmetropic, and 9% were hyperopic. The common ancestor (proband) is represented multiple times (eight times) by a small black pointer. Curved lines show repeat instances of the same dog(s) within the pedigree.
Figure 2.
 
Distribution of refractive error (mean SER of both eyes) within the sample of 116 dogs who were all descendents of one stud dog with myopia of −2.00 D. Histogram bin width, 0.50 D. Mean refractive error of this group = −0.41 D (range, −5.38 to +1.65 D).The bimodal nature of this distribution suggests that dogs with myopia tended to form a distinct group, with the major group having a near-normal distribution about low hyperopia.
Figure 2.
 
Distribution of refractive error (mean SER of both eyes) within the sample of 116 dogs who were all descendents of one stud dog with myopia of −2.00 D. Histogram bin width, 0.50 D. Mean refractive error of this group = −0.41 D (range, −5.38 to +1.65 D).The bimodal nature of this distribution suggests that dogs with myopia tended to form a distinct group, with the major group having a near-normal distribution about low hyperopia.
Figure 3.
 
Plot of refractive error (mean SER of both eyes) versus age for the 116 dogs in the sample. All dogs had ages within the range 1 to 8 years. Shown is the best fitting (least squares) linear fit to the data (SER = 0.003(age) + 0.42). There was no significant relationship between refractive error and age for the dogs in this sample (Pearson R = 0.0047, P = 0.52).
Figure 3.
 
Plot of refractive error (mean SER of both eyes) versus age for the 116 dogs in the sample. All dogs had ages within the range 1 to 8 years. Shown is the best fitting (least squares) linear fit to the data (SER = 0.003(age) + 0.42). There was no significant relationship between refractive error and age for the dogs in this sample (Pearson R = 0.0047, P = 0.52).
Table 1.
 
Status of Dogs in the Family of Labrador Retrievers Studied, at the Time of Analysis
Table 1.
 
Status of Dogs in the Family of Labrador Retrievers Studied, at the Time of Analysis
Total refracted 116
Outside age range (<1 or >8 y) 131
In New Zealand but not available for testing 71
Outside New Zealand 40
Ocular disease 24
Cross-breed 14
Deceased 11
Total family 407
Table 2.
 
Models Used to Test for Statistical Significance of Variables
Table 2.
 
Models Used to Test for Statistical Significance of Variables
Model Comment Log Likelihood χ2 P
G + S + C + L + E Full model −66.50 N/A N/A
S + C + L + E Test for additive genetic effect −74.14 15.28 9.3× 10−5 *
G + C + L + E Testing for effect of sex −67.31 1.62 0.20
G + S + L + E Testing for effect of coat color −67.56 2.12 0.15
G + S + C + E Testing for effect of litter size −71.64 10.28 0.0013*
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×