SEARCH

SEARCH BY CITATION

Keywords:

  • quantitative trait locus;
  • osteoporosis;
  • genetics;
  • hip fracture;
  • femoral structure

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Femoral structure contributes to bone strength at the proximal femur and predicts hip fracture risk independently of bone mass. Quantitative components of femoral structure are highly heritable traits. To identify genetic loci underlying variation in these structural phenotypes, we conducted an autosomal genome screen in 309 white sister pairs. Seven structural variables were measured from femoral radiographs and used in multipoint sib-pair linkage analyses. Three chromosomal regions were identified with significant evidence of linkage (log10 of the odds ratio [LOD] > 3.6) to at least one femoral structure phenotype. The maximum LOD score of 4.3 was obtained for femur neck axis length on chromosome 5q. Evidence of linkage to chromosome 4q was found with both femur neck axis length (LOD = 3.9) and midfemur width (LOD = 3.5). Significant evidence of linkage also was found to chromosome 17q, with a LOD score of 3.6 for femur head width. Two additional chromosomal regions 3q and 19p gave suggestive (LOD > 2.2) evidence of linkage with at least two of the structure phenotypes. Chromosome 3 showed evidence of linkage with pelvic axis length (LOD = 3.1), midfemur width (LOD = 2.8), and femur head width (LOD = 2.3), spanning a broad (60 cm) region of chromosome 3q. Linkage to chromosome 19 was supported by two phenotypes, femur neck axis length (LOD = 2.8) and femur head width (LOD = 2.8). This study is the first genome screen for loci underlying variation in femoral structure and represents an important step toward identifying genes contributing to the risk of osteoporotic hip fracture in the general population.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

AGE-RELATED OSTEOPOROTIC fractures of the hip are a major public health problem, occurring at a rate of over 1.3 million per year in the United States and accounting for over $10 billion in healthcare expenses.(1) Hip fracture is a stochastic event resulting from recurrent trauma, usually from falls, and the loss of bone strength at the proximal femur. Clinically measurable components of bone strength include bone mass and bone structure. Femoral structure has been shown to predict hip fracture risk.(2–6) Several studies have shown that the longer the hip axis, the greater the risk of hip fracture.(3–5) Furthermore, despite lower mean bone mass, Asians have lower hip fracture rates than whites, perhaps partly because of the protective effect of a shorter mean hip axis length (HAL).(7–10)

Genetic factors play a major role in determining femoral structure. Twin and family studies have indicated significant heritability of HAL(11, 12) and other structural phenotypes.(13) Race is also an important determinant of HAL, with both Asians(7–9) and blacks(8, 14–16) having shorter HAL than whites.

Several genes likely contribute to the observed genetic component of variability in femoral structure. One method of identifying these genes is the candidate gene approach, in which known genes hypothesized to contribute to the phenotype of interest are investigated for possible effects on the phenotype. A study using this approach in twins failed to detect an effect of genetic polymorphism at the vitamin D receptor (VDR) on HAL.(17)

In an alternative approach, we have used highly polymorphic markers distributed throughout the genome to detect chromosomal regions harboring known or novel genes influencing the variability in femoral structure. We report the results of an autosomal genomewide linkage screen with seven femoral structure variables measured from radiographs in 309 healthy white, premenopausal sister pairs. These linkage results are an important first step in localizing and identifying the genes influencing femoral structure.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Subjects

Sibling pairs consisting of 481 healthy white premenopausal sisters aged 20-45 years from 221 families were recruited in Indiana(18) (Table 1). These 481 sisters are a subset of the 636 sisters previously analyzed to detect loci contributing to variability in bone mineral density (BMD). Radiographs were added to the protocol in the second year of the study. Blood samples were obtained from all participating sisters and from one of their parents. DNA was isolated using standard techniques.(19) Height and weight, along with oral contraceptive and smoking (pack-years) history were obtained. Informed consent (Indiana University Institutional Review Board approval: 8502-8523) was obtained from all participants.

Table Table 1.. Sample Pedigree Characteristics
Thumbnail image of

Phenotypic measurements

Radiographs of the lower pelvis to include both upper femurs in 15° internal rotation were taken on standard X-ray equipment using a focal distance of 40 in. (100 cm) and PDG1 film (Eastman Kodak Corp., Rochester, NY, USA). The thickness of the calcar femorale at the lesser trochanter, the width of the femur and its medulla at the midshaft, the length of the femur neck and pelvic axes, and maximal width of the femur head and minimal width of the femur neck (Fig. 1) were measured directly from the radiographs by one observer (G.L.) using a Digimatic Caliper (Mitutoyo Corp., Kawasaki, Japan) as previously described.(20) Total body fat, lean body mass, and areal BMD of the femoral neck were measured by dual-energy X-ray absorptiometry (DXA; Lunar DPXL; Lunar Corp., Madison, WI, USA).

thumbnail image

Figure FIG. 1.. (A) Femoral structure measurements shown schematically. NW, femur neck width; HW, femur head width. (B) A representative radiograph of the upper femur, from which the structure measurements were made.

Download figure to PowerPoint

Marker genotyping

A total of 270 markers were genotyped for 309 white sister pairs and one of their parents, if available. Markers were selected from the ABI/Prism genome screening set (version 1.0; Applied Biosystems Inc., Foster City, CA, USA), with 44 additional public markers selected to fill large intermarker gaps. All markers were dinucleotide repeat polymorphisms and were highly informative, with heterozygosities greater than 0.70. Polymerase chain reactions for each marker were performed separately and products were combined before gel electrophoresis. Data were collected using the 373A automated DNA sequencer (Applied Biosystems Inc.) and genotyped using the Genescan 672 and Genotyper software (Applied Biosystems Inc.) The average spacing of the markers was 12.9 cm.

Statistical analysis

Stepwise regression analysis was used to identify significant non-BMD covariates with the femoral structure variables. Regression residuals, representing covariate-adjusted femoral structure measures, were computed and used in all subsequent analyses. To prioritize phenotypes for subsequent linkage analyses, approximate heritabilities for each of the femoral structure measures were calculated from the full-sibling data. Broad-sense heritability (H2) was estimated as twice the sibling intraclass correlation, according to the method of Falconer.(21) To avoid weighting the larger families excessively in the heritability calculation, we used the correction of Donner and Koval(22) for sibships larger than two. The marker genotype data were used to verify the full-sibling relationships among the subjects using the computer programs RELATIVE(23) and RELPAIR.(24) Five half-sibling pairs were eliminated from further analyses because of significantly lower sharing of marker alleles identical by descent (IBD) than would be expected for full siblings.

Multipoint quantitative linkage analysis was performed for each femoral structure phenotype using the maximum likelihood variance estimation method as implemented in the computer package Mapmaker/SIBS.(25) Log10 of the odds ratio (LOD) scores were computed at 1-cm intervals along each autosome using all possible sibling pairs from families that had more than two sisters. To confirm the robustness of linkage findings, analyses also were performed using only independent sibling pairs and implementing the more conservative Haseman-Elston regression approach.(25) Observed allele frequencies in the individuals genotyped for the genome screen were used. Marker order and map positions were obtained from the Marshfield electronic database (http://www.marshmed.org/genetics/).

All statistical tests of linkage were based on IBD marker allele sharing. Alleles are IBD if siblings inherit the same marker allele from the same parent. If the marker being tested is in close physical proximity to a gene influencing femoral structure, then siblings with similar phenotypic values would be expected to share marker alleles IBD. Conversely, siblings with dissimilar femoral structure traits would be expected to share fewer marker alleles IBD near the gene influencing the phenotype. An advantage of quantitative linkage methods as used here is that no arbitrary cut-off for “high” or “low” phenotypic values is necessary; therefore, all sibling pairs are included in the analysis.

We performed pairwise correlation analysis between the femoral structure phenotypes to aid in the interpretation of the linkage results. These analyses are especially useful if evidence of linkage is observed in the same chromosomal region with multiple phenotypes. Because correlations between BMD at the femoral neck and the structural variables have been reported,(5, 20) we also included femoral neck areal BMD in our correlation analysis. Linkage results for the BMD phenotypes already have been reported.(26)

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Four hundred eighty-one sisters in 221 sibships were ascertained (Table 1). Genome screen marker genotypes, covariate measurements, and femoral structure phenotypes were completed on all sisters except for six on whom the radiograph did not permit measurement of midfemur width. The mean age of the sisters was 34.1 years, and the mean difference in age between sisters was 3.6 years. Examination of histograms and normal probability plots for all of the femoral structure measures and covariates revealed no substantial deviations from normality, apart from a mild skewness of the distribution of body weight. Of the covariates studied, only age, height, weight, and lean body mass approached significance (p < 0.10) in stepwise model fitting. Residuals from regression model fitting with these four measures as independent variables were used as covariate-adjusted phenotypic values in all subsequent linkage analyses. Heritabilities for each of the femoral structure phenotypes, along with the number of sister pairs available for each phenotype, are shown in Table 2. The observed heritabilities are high, with genetic factors accounting for 60-80% of the variability in the femoral structure phenotypes, making these phenotypes excellent candidates for genetic linkage analyses.

Table Table 2.. Simple Statistics, Sample Size and Heritability (H2) Estimates for Structural Phenotypes at the Proximal Femur
Thumbnail image of

Multipoint linkage analysis of the seven femoral structure phenotypes identified three chromosomal regions with significant evidence of linkage (LOD > 3.6) to at least one measure (Table 3), according to the widely accepted criteria of Lander and Kruglyak(27) for genome scans of sibling data. The maximum LOD score attained in the genome screen was 4.3 on chromosome 5q, near the marker D5S647 with femur neck axis length (Fig. 2A). Other regions with significant evidence of linkage to particular structural measures were chromosome 4q, with a LOD score of 3.9 for femur axis length near the marker D4S428 (Fig. 2B), and chromosome 17q, with a LOD score of 3.6 for femur head width near the marker D17S791 (Fig. 2C). Four additional chromosomal regions (3q, 19p, 9q, and 7q) gave suggestive evidence of linkage (LOD > 2.2) with at least one of the measures (Table 3).

Table Table 3.. Summary of Suggestive and Significant Genome Screen Results for Femoral Structure Phenotypes
Thumbnail image of
thumbnail image

Figure FIG. 2.. LOD score plots showing all markers used in the genome screen for chromosomes (A) 5, (B) 4, (C) 17, (D) 19, and (E) 3. Genetic markers are indicated by D-name along the top of each plot; marker positions are shown on the x axis. Structural phenotypes are as follows: femur head width ([2 point thick rule 2 picas long]), pelvic axis length (— —), midfemur shaft width (- -), and femur neck axis length (- - -).

Download figure to PowerPoint

In a number of instances, several phenotypes were linked to the same chromosomal region and, in some cases, with the maximum LOD score occurring at the same genetic marker. Chromosome 4q showed linkage with femur neck axis length (LOD = 3.9) and with the midfemur shaft width (LOD = 3.5) in the same chromosomal region, with the two peaks less than 15 cm apart (Fig. 2B). Similarly, the suggestive linkage on chromosome 19q with femur neck axis length (LOD = 2.8) is corroborated by linkage to femur head width (LOD = 2.8), with the peaks occurring within 2 cm of one another (Fig. 2D). Chromosome 3q shows suggestive evidence of linkage with three different phenotypes: pelvic axis length (LOD = 3.1), midfemur shaft width (LOD = 2.8), and femur head width (LOD = 2.3). These linkage peaks occur within a large (∼60 cm) span of chromosome 3q between the markers D3S1271 and D3S1614 (Fig. 2E). Evidence of linkage to the same chromosomal regions (5q, 4q, 17q, 3q, 19p, 9q, and 7q) was observed both when limiting the analysis to only the independent sibling pairs in the data set and when using the alternate, more conservative Haseman-Elston regression method(25) to test for linkage.

Results of the pairwise correlation analysis of the femoral structure phenotypes and BMD at hip are shown in Table 4. Several of the pairs of structure measures, including shaft width with neck axis length and head width with both neck width and neck axis length, were correlated moderately, with Pearson correlation coefficients between 0.4 and 0.6. However, correlations were 0.25 or below between BMD and the structural phenotypes.

Table Table 4.. Phenotypic Correlations for Femoral Structure Phenotypesa
Thumbnail image of

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Our study reports high heritability of femoral structure phenotypes, consistent with previous twin-based estimates of the heritability of HAL.(11, 12) We conducted the first genome scan for femoral structure phenotypes in a sample of healthy white sister pairs and found evidence for linkage to several chromosomal regions. The strongest evidence for linkage in our sample was to chromosome 5q with femur neck axis length (LOD = 4.3). Three of the linkage findings (on chromosomes 4, 5, and 17) surpass the criterion (LOD > 3.6) of Lander and Kruglyak(27) for genomewide significance. We also have identified four other chromosomal regions with suggestive evidence of linkage (LOD > 2.2) to one or more femoral structure phenotypes.

The greatest evidence for linkage observed in our sample was to chromosome 5q with femur axis length (LOD = 4.3). The region implicated by our linkage findings (D5S647-D5S644) contains the gene for cartilage linking protein 1, believed to be important for the formation of proteoglycan aggregates and normal organization of hypertrophic chondrocytes.(28) Evidence for a gene on chromosome 4p influencing femoral structure is supported by our linkage results both with femur neck axis length and femur midshaft width (LOD = 3.5). The maximum LOD scores for these two phenotypes are within 2 cm of one another, well within the limit of resolution of the mapping methods used.(29) The linkage peaks for both femur neck axis length and femur midshaft width also are within 15 cm of an attractive candidate gene, the bone morphogenic protein 3 (BMP-3) gene, thought to be involved in bone morphogenesis and cell differentiation.(30, 31)

The region of chromosome 17 where we find significant linkage with femur head width (LOD = 3.6) contains several interesting candidate genes. These include type 1 collagen (COL1A1); chondroadherin, thought to be involved in cartilage formation and maintenance(32); and a cluster of homeobox (HOX) genes. HOX genes have been implicated in the control of patterns of growth and body structure in numerous animal systems.(33)

Additionally, we find several other chromosomal regions with suggestive evidence of linkage to one or more of the structural phenotypes. The broad region of linkage on chromosome 3q shows evidence of linkage with three structural phenotypes (LOD = 3.12 with pelvic axis length, LOD = 2.8 with femur midshaft width, and LOD = 2.3 with femur head width). This region contains two attractive candidate genes: type 8 collagen (COL8A1) and procollagen-lysine oxyglutamate dioxygenase 2 (PLOD2), which is believed to be involved in tissue-specific collagen cross-linking patterns.(34) Linkage is observed on chromosome 19 with two different structure phenotypes: LOD = 2.8 with femur neck axis length and LOD = 2.82 with femur head width. This linked region is near the structural gene for cartilage oligomeric matrix protein (COMP),(35) a calcium-binding protein believed to be involved in the cross-linking of types I, II, and IX collagen.

We previously reported linkage of femoral neck and lumbar spine BMD to chromosomes 1p, 5q, 6p, 11q, and 22q(18, 26) in a sample consisting of the subjects in the current report and an additional 120 sibling pairs unique to the BMD analysis. In this study, as we have shown in other populations,(20) femoral neck BMD is correlated only moderately with structural phenotypes (r ≤ 0.25), particularly calcar width and total femur midshaft width. For this reason, it is perhaps not surprising that we observe little or no overlap in the linked chromosomal regions for femoral structure and BMD at the hip and spine. This suggests that there is little overlap between the set of genes regulating BMD and those regulating bone structure.

The region of chromosome 5q near marker D5S647 showing significant linkage to femur neck axis length is a substantial distance (90 cm) from the region near D5S422 where we observed linkage with femoral neck BMD. The resolution or the ability to estimate the position of a genetic locus using linkage analysis is probably limited to 20 cm or 30 cm.(29) However, even with these wide CIs, our linkage findings with femur neck axis length and with neck BMD on chromosome 5q are unlikely to represent the pleiotropic action of a single locus.

On the other hand, we do observe evidence of linkage to several chromosomal regions (3, 4, and 19) with multiple femoral structure phenotypes. The hypothesis that a single locus or group of linked loci has pleiotropic effects is supported by the linkage of both the femur neck axis length and the femur midshaft width to the same region of chromosome 4 and of femur neck axis length and femur head width to the same region of chromosome 19q. It is interesting that these two phenotypes have among the highest correlation of the femoral structure phenotypes (Table 4, r = 0.40 and r = 0.43, respectively). The linkage to chromosome 3q is observed with both pelvic axis length and femur midshaft width; however, these two phenotypes are not significantly correlated (r = 0.02) in our sample. This may indicate the presence of two loci in this region of chromosome 3, one primarily affecting pelvic axis length and the other femur midshaft width or, alternatively, a single locus with a small effect on each of the phenotypes.

The structural phenotypes at the proximal femur that contribute to bone strength and/or predict hip fracture have not been extensively examined. Those measured in this study have been suggested previously to be important.(4, 6) Because it is readily measured from DXA output, HAL is perhaps the most widely studied structural measure. However, HAL is a composite measurement of femur neck axis and pelvic axis lengths, neither of which can be measured easily from DXA printouts. They require better visualizing techniques such as radiography and computed tomography for precise measurement. In this study, femur neck axis length, which is likely to contribute substantially to bone strength at the hip, was found to be linked to three chromosomal regions. Pelvic axis length, which probably contributes little to overall bone strength at the hip, was found to be linked to a locus different from any of those linked to femur neck axis length.

In summary, we report linkage of femoral structure phenotypes to several broad chromosomal regions. These regions should be confirmed by increasing the number of sisters studied and by saturation mapping with additional markers. An enlarged sample along with increased marker density should enable us to confirm the linkage findings reported here, as well as map the loci with greater accuracy for subsequent positional cloning. In addition, a larger sample size will increase the power to detect additional loci having smaller effect, thus allowing epistatic effects to be examined. Femoral structure is an important component of bone strength at the hip and a predictor of fracture risk. Elucidation of the genes underlying the normal variation in femoral structure is likely to increase basic knowledge of the biochemical pathways involved in skeletal biology, to aid in developing tests to identify patients at risk of hip fracture, and to contribute to the development of therapeutic interventions to prevent and treat osteoporotic fractures of the hip.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

We gratefully acknowledge the sisters and parents who participated in this study, as well as the study coordinators, without whom this work could not have been accomplished. Support from PHS RO1AR43476, MOI00750, and Medical and Molecular Genetics training grant PHS T32 HD07373 is gratefully acknowledged. This research was funded in part by collaboration between Hoffman La-Roche, Inc./Boehringer Mannheim and Axys Pharmaceuticals, Inc.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Osteoporosis Prevention, Diagnosis, and Therapy. NIH Consensus Statement 17[2], March 27, 2000, pp. 136.
  • 2
    Horsman A, Marshall DH, Peacock M 1985 A stochastic model of age-related bone loss and fractures. Clin Orthop 195:207215.
  • 3
    Peacock M, Turner CH, Liu G, Manatunga AK, Timmerman L, Johnston CC Jr 1995 Better discrimination of hip fracture using bone density, geometry and architecture. Osteoporos Int 5:167173.
  • 4
    Faulkner KG, Cummings SR, Nevitt MC, Pressman A, Jergas M, Genant HK 1995 Hip axis length and osteoporotic fractures. Study of Osteoporotic Fractures Research Group. J Bone Miner Res 10:506508.
  • 5
    Karlsson KM, Sernbo I, Obrant KJ, Redlund-Johnell I, Johnell O 1996 Femoral neck geometry and radiographic signs of osteoporosis as predictors of hip fracture. Bone 18:327330.
  • 6
    Glüer CC, Cummings SR, Pressman A, Li J, Glüer K, Faulkner KG, Grampp S, Genant HK 1994 Prediction of hip fractures from pelvic radiographs: The study of osteoporotic fractures. The Study of Osteoporotic Fractures Research Group. J Bone Miner Res 9:671677.
  • 7
    Chin K, Evans MC, Cornish J, Cundy T, Reid IR 1997 Differences in hip axis and femoral neck length in premenopausal women of Polynesian, Asian and European origin. Osteoporos Int 7:344347.
  • 8
    Cummings SR, Cauley JA, Palermo L, Ross PD, Wasnich RD, Black D, Faulkner KG 1994 Racial differences in hip axis lengths might explain racial differences in rates of hip fracture. Study of Osteoporotic Fractures Research Group. Osteoporos Int 4:226229.
  • 9
    Nakamura T, Turner CH, Yoshikawa T, Slemenda CW, Peacock M, Burr DB, Mizuno Y, Orimo H, Ouchi Y, Johnston CC 1994 Do variations in hip geometry explain differences in hip fracture risk between Japanese and white Americans? J Bone Miner Res 9:10711076.
  • 10
    Farmer ME, White LR, Brody JA, Bailey KR 1984 Race and sex differences in hip fracture incidence. Am J Public Health 74:13741380.
  • 11
    Arden NK, Baker J, Hogg C, Baan K, Spector TD 1996 The heritability of bone mineral density, ultrasound of the calcaneus and hip axis length: A study of postmenopausal twins. J Bone Miner Res 11:530534.
  • 12
    Flicker L, Faulkner KG, Hopper JL, Green RM, Kaymacki B, Nowson CA, Young D, Wark JD 1996 Determinants of hip axis length in women aged 10-89 years: A twin study. Bone 18:4145.
  • 13
    Koller DL, Peacock M, Conneally PM, Liu G, Hui SL, McClintock R, Christian JC, Johnston CC, Econs MJ, Foroud T 1998 Heritability of bone-related phenotypes in Caucasian and African-American sister pairs. Am J Hum Genet 63:A214.(abstract)
  • 14
    Mikhail MB, Vaswani AN, Aloia JF 1996 Racial differences in femoral dimensions and their relation to hip fracture. Osteoporos Int 6:2224.
  • 15
    Daniels ED, Pettifor JM, Schnitzler CM, Moodley GP, Zachen D 1997 Differences in mineral homeostasis, volumetric bone mass and femoral neck axis length in black and white South African women. Osteoporos Int 7:105112.
  • 16
    Theobald TM, Cauley JA, Gluer CC, Bunker CH, Ukoli FA, Genant HK 1998 Black-white differences in hip geometry. Study of Osteoporotic Fractures Research Group. Osteoporos Int 8:6167.
  • 17
    Arden NK, Keen RW, Lanchbury JS, Spector TD 1996 Polymorphisms of the vitamin D receptor gene do not predict quantitative ultrasound of the calcaneus or hip axis length. Osteoporos Int 6:334337.
  • 18
    Koller DL, Rodriguez LA, Christian JC, Slemenda CW, Econs MJ, Hui SL, Morin P, Conneally PM, Joslyn G, Curran ME, Peacock M, Johnston CC, Foroud T 1998 Linkage of a QTL contributing to normal variation in bone mineral density to chromosome 11q12-13. J Bone Miner Res 13:19031908.
  • 19
    Madisen L, Hoar DI, Holroyd CD, Crisp M, Hodes ME 1987 DNA banking: The effects of storage of blood and isolated DNA on the integrity of DNA. Am J Med Genet 27:379390.
  • 20
    Peacock M, Liu G, Carey M, Ambrosius W, Turner CH, Hui S, Johnston CC 1998 Bone mass and structure at the hip in men and women over the age of 60 years. Osteoporos Int 8:231239.
  • 21
    Falconer DS 1998 Introduction to Quantitative Genetics, 3rd ed. Longman Scientific and Technical, Essex, UK, pp. 163176.
  • 22
    Donner A, Koval JJ 1980 The estimation of intraclass correlation in the analysis of family data. Biometrics 36:1925.
  • 23
    Goring HH, Ott J 1997 Relationship estimation in affected sib pair analysis of late-onset diseases. Eur J Hum Genet 5:6977.
  • 24
    Boehnke M, Cox NJ 1997 Accurate inference of relationships in sib-pair linkage studies. Am J Hum Genet 61:423429.
  • 25
    Kruglyak L, Lander ES 1995 Complete multipoint sib-pair analysis of qualitative and quantitative traits. Am J Hum Genet 57:439454.
  • 26
    Koller DL, Econs MJ, Morin PA, Christian JC, Hui SL, Parry P, Curran ME, Rodriguez LA, Conneally PM, Joslyn G, Peacock M, Johnston CC, Foroud T 2000 Genome screen for QTLs contributing to normal variation in bone mineral density and osteoporosis. J Clin Endocrinol Metab 85:31163120.
  • 27
    Lander E, Kruglyak L 1995 Genetic dissection of complex traits: Guidelines for interpreting and reporting linkage results. Nat Genet 11:241247.
  • 28
    Watanabe H, Yamada Y 1999 Mice lacking link protein develop dwarfism and craniofacial abnormalities. Nat Genet 21:225229.
  • 29
    Roberts SB, MacLean CJ, Neale MC, Eaves LJ, Kendler KS 1999 Replication of linkage studies of complex traits: An examination of variation in location estimates. Am J Hum Genet 65:876884.
  • 30
    Aberg T, Wozney J, Thesleff I 1997 Expression patterns of bone morphogenetic proteins (BMPs) in the developing mouse tooth suggest roles in morphogenesis and cell differentiation. Dev Dyn 210:383396.
  • 31
    Amedee J, Bareille R, Rouais F, Cunningham N, Reddi H, Harmand MF 1994 Osteogenin (bone morphogenic protein 3) inhibits proliferation and stimulates differentiation of osteoprogenitors in human bone marrow. Differentiation 58:157164.
  • 32
    Mizuno M, Fujisawa R, Kuboki Y 1996 Bone chondroadherin promotes attachment of osteoblastic cells to solid-state substrates and shows affinity to collagen. Calcif Tissue Int 59:163167.
  • 33
    Boncinelli E 1999 Homeobox genes in development. Adv Neurol 79:8194.
  • 34
    Uzawa K, Grzesik WJ, Nishiura T, Kuznetsov SA, Robey PG, Brenner DA, Yamauchi M 1999 Differential expression of human lysyl hydroxylase genes, lysine hydroxylation, and cross-linking of type I collagen during osteoblastic differentiation in vitro. J Bone Miner Res 14:12721280.
  • 35
    Thur J, Rosenberg K, Nitsche DP, Pihlajamaa T, Ala-Kokko L, Heinegard D, Paulsson M, Maurer P 2001 Mutations in cartilage oligomeric matrix protein causing pseudoachondroplasia and multiple epiphyseal dysplasia affect binding of calcium and collagen I, II and IX. J Biol Chem 276:60836092.