1 3
Hum Genet (2014) 133:755–768DOI 10.1007/s00439-013-1409-x
OrIGInal InvestIGatIOn
Identification of rare genetic variants in novel loci associated with Paget’s disease of bone
Mariejka Beauregard · Edith Gagnon · Sabrina Guay‑Bélanger · Jean Morissette · Jacques P. Brown · Laëtitia Michou
received: 15 september 2013 / accepted: 8 December 2013 / Published online: 27 December 2013 © springer-verlag Berlin Heidelberg 2013
15q24 locus. We located 71 of these 126 rare variants in an intron, 30 in an exon and 9 in an untranslated region. 60 % of these variants were located in functionally rel-evant gene regions. among the 12 missense rare variants in PDB, two (rs62620995 in TM7SF4; rs62641691 in CD276) were predicted to be damaging by in silico anal-ysis tools. rs62620995, which altered a conserved amino acid (p.leu397Phe) in the TM7SF4 gene, encoding the DC-staMP protein involved in osteoclastogenesis through ranK signaling pathway, was found to have a marginal association with PDB (p = 0.09). rs35500845, located in the CTHRC1 gene, which encodes a regulator of collagen matrix deposition, was also associated with PDB in the French-Canadian population (p = 0.046).
Introduction
Paget’s disease of bone (PDB [MIM 602080]) is the second most common metabolic bone disorder after osteoporosis (roodman and Windle 2005). It is characterized by a focal increase in bone remodeling resulting in abnormal bone architecture (lyles et al. 2001). the disease remains local-ized to the affected bones and does not spread to adjacent bones (roodman and Windle 2005). PDB has an autosomal dominant mode of inheritance in about one-third of cases (Morales-Piga et al. 1995; siris et al. 1991). Disease-caus-ing mutations have been identified in a single gene to date, the SQSTM1 gene. approximately one-third of patients with familial form of PDB are carrier of a SQSTM1 muta-tion (Yan Jenny Chung and van Hul 2011). recently, five novel loci associated with PDB have been identified by two genome-wide association studies (GWas) conducted in PDB patients without SQSTM1 mutations (supplemen-tary table 1) (albagha et al. 2010, 2011). the cumulative
Abstract In genome-wide association studies, single nucleotide polymorphisms located in five novel loci were associated with PDB. We aimed at identifying rare genetic variants of candidate genes located in these loci and search for genetic association with PDB in the French-Canadian population. exons, promoter and exon–intron junctions from patients with familial PDB and healthy individu-als were sequenced in candidate genes, located within novel loci associated with PDB in our population. rare variant was defined by a minor allele frequency <0.05 or absent from dbsnP (nCBI). We sequenced seven genes in 1p13 locus, three genes in 7q33, three genes in 8q22, and five genes in 15q24 locus. We identified 126 rare vari-ants in at least one patient with PDB of whom 55 were located in 1p13 locus, 32 in 7q33, 10 in 8q22 and 29 in
Accession numbers nucleotide sequence data for novel reported rare variants (table 2) are permanently available from the ena browser at http://www.ebi.ac.uk/ena/data/view/HG005313-HG005349 once they are released into the public domain, in the ‘european nucleotide archive’ database.
Electronic supplementary material the online version of this article (doi:10.1007/s00439-013-1409-x) contains supplementary material, which is available to authorized users.
M. Beauregard · e. Gagnon · s. Guay-Bélanger · J. Morissette · J. P. Brown · l. Michou CHU de Québec research Centre, Quebec, QC, Canada
M. Beauregard · s. Guay-Bélanger · J. P. Brown · l. Michou Division of rheumatology, Department of Medicine, laval University, Quebec, QC, Canada
J. P. Brown · l. Michou (*) Department of rhumatologie-s763, CHU de Québec, 2705 boulevard laurier, Quebec, QC G1v 4G2, Canadae-mail: [email protected]
756 Hum Genet (2014) 133:755–768
1 3
population-attributable risk of these five loci in addition with two already known loci was estimated at 86 % among non-carriers of an SQSTM1 mutation. the risk of devel-oping PDB increases with the increasing number of risk alleles carried (albagha et al. 2011).
the common diseases–rare variants hypothesis sug-gests that a significant part of the genetic susceptibility to common diseases is caused by multiple variants with minor allele frequency (MaF) between those of muta-tions (<1 %) and common variants (>5 %) (Bodmer and Bonilla 2008; Frazer et al. 2009; Gorlov et al. 2008). these rare variants are independent of each other and can confer a detectable risk of developing a disease (Bodmer and Bonilla 2008). rare variants may have been respon-sible for some associations between a disease and a com-mon variant through GWas (Cirulli and Goldstein 2010; Gibson 2011; Manolio et al. 2009). the French-Canadian population is appropriate for rare genetic variants research. Because of its founder effect, it is potentially enriched in rare variants and benefits from greater statistical power (Bodmer and tomlinson 2010; Marian 2012). large fami-lies (10–20 siblings until recently), and religious records preserved from first arrivals also facilitate genetic studies in this population.
Our research hypothesis states that rare genetic variants located in novel PDB-associated loci could play a role in genetic susceptibility to PDB. In this study, we aimed to identify rare genetic variants in candidate genes for PDB, located nearby common variants associated with PDB, to identify a genetic association of one or more of these rare variants with PDB and to develop an effective method for exploring the role of rare variants in genetic susceptibility to a disease.
Materials and methods
Individuals
after information, all individuals signed a consent form. the research project was approved by the ethics commit-tee of the Centre Hospitalier de l’Université laval (CHUl). all cases selected for this project were examined by Dr. Jacques Brown or Dr. laëtitia Michou, rheumatologists at the CHU de Québec. a full assessment of bone health, including measurement of total serum alkaline phosphatase, bone X-rays of the pelvis and skull and total bone scan was performed for each patient. the criteria used to diagnose PDB included (1) a typical aspect of PDB on bone X-rays and/or (2) an abnormal bone scan. Controls were healthy individuals without any known personal or familial his-tory of PDB and with normal alkaline phosphatase levels
at inclusion. none of the controls carried any mutation in SQSTM1 gene. Healthy individuals were not matched for age and sex with PDB patients. 58.1 % of patients were male with mean age at inclusion of 62.5 ± 10.9 years, and 28.2 % of controls were male with a mean age at inclusion of 64.7 ± 10.9 years. all individuals studied were from the French-Canadian population. the Dna of each individual was extracted from peripheral blood samples by a standard procedure.
Genotyping of snPs in novel loci found to be associated with PDB in the literature
Genotyping of these snPs (i.e., rs484959, rs499345, rs10494112, rs4294134, rs2458413, rs10498635 and rs5742915) relied on sequenom MassarraY snP Multi-plex technology, performed at the Plateforme de séquen-çage et de génotypage des génomes du Centre de recherche du CHU de Québec, in 240 unrelated PDB patients (includ-ing 23 patients carrier of a SQSTM1/P392L mutation) and 297 unrelated healthy controls from the French-Canadian population. Briefly, purified Dna solution containing mul-tiplexed primer-based extension reaction (iPleX reaction) products was dispensed from the 384-well microplate onto a 384-pad silicon microchip using the MassarraY nan-odispenser. the mass of each snP allele was detected on the MassarraY Compact MalDI-tOF (matrix-assisted laser desorption/ionization-time of flight) mass spectrome-ter, and the results were analyzed with MassarraY typer software. Duplicated samples were included to verify the allele calls.
selection of candidate genes
all genes located in an interval of 0.5–1 Mb upstream and downstream of associated common variants in our popula-tion (defined by an uncorrected p < 0.05 in allelic or geno-typic association study) were identified through Geneloc (http://genecards.weizmann.ac.il/geneloc/index.shtml). the length of the interval was determined according to gene density of the locus. Genes containing or located near the common variants associated with PDB were all selected. Other genes studied in the project were chosen because of their potential role in the pathogenesis of PDB according to their known functions. Database EntrezGene (http://www.ncbi.nlm.nih.gov/gene/), Genatlas (http://www.genatlas.org/), GeneCards (http://www.genecards.org/), Online Mendelian Inheritance in Man (OMIM) (http://www.ncbi.nlm.nih.gov/omim/), UniProtKB (http://www.uniprot.org/help/uniprotkb) and WikiGenes (http://www.wikigenes.org/) were consulted in order to select the best candidate genes for PDB.
757Hum Genet (2014) 133:755–768
1 3
search for rare variants in candidate genes by bidirectional sequencing
We searched for rare variants in a “discovery sample” which included 32 patients with a familial form of PDB and 4 healthy individuals from the French-Canadian popu-lation. the number of patients carrier of SQSTM1/P392L mutation included in the discovery set has varied from 0 to 16 patients. each patient with a familial form belonged to a different family to avoid bias of relatedness. all coding exons, exon–intron junctions, 5′Utr and 3′Utr regions of selected genes were sequenced. Promoter regions, esti-mated to be comprised in the 500 bp preceding the atG of each gene, were also sequenced. Genes were first amplified by polymerase chain reaction (PCr). the oligonucleotide primers were designed using Primer 3 (http://biotools.umassmed.edu/bioapps/primer3_www.cgi). PCr products size was verified by electrophoresis on agarose gel. the ampli-fication products corresponding to the expected size were purified. Both strands were sequenced at the Plateforme de séquençage et de génotypage des génomes du Centre de recherche du CHU de Québec with elongation termina-tors Big Dye Deoxy Terminator v 3.1 cycle (applied Bio-systems) on an aBI 3730xl sequencer. the Dna sequences were analyzed with staDen software version 2.0.0b8 in comparison with the reference sequences (staden 1996). We looked at the chromatograms for every deviation from the reference sequences. all variants carried by at least one individual for which the minor allele frequency (MaF) was below 5 % (0.05) according to the database EntrezSNP (http://www.ncbi.nlm.nih.gov/gene/) were considered as rare variants. variants for which the MaF was unavailable in EntrezSNP as well as variants not listed in EntrezSNP were also considered as rare variants.
In silico characterisation of rare genetic variants predicted function
In silico predictions from PolyPhen (http://genetics.bwh.harvard.edu/pph2/), sIFt (http://sift.jcvi.org/) and Con-del (http://bg.upf.edu/condel/home) were obtained for all variants that altered an amino acid. the type of amino acid change (radical versus conservative) was determined according to the criteria set published by Zhang (2000). Conservation of amino acids through evolution was assessed using COBalt (http://www.ncbi.nlm.nih.gov/tools/cobalt/). Known domains of the proteins encoded by the studied genes were identified through database Uni-ProtKB (http://www.uniprot.org/help/uniprotkb) and Pfam (http://pfam.sanger.ac.uk/). splice sites alteration was assessed using Human splicing Finder (http://www.umd.be/HsF/). alteration of binding sites of transcription factors involved in bone regulation (supplementary table 2) was
evaluated with tFsearCH (http://www.cbrc.jp/research/db/tFsearCH.html) and Consite (http://asp.ii.uib.no:8090/cgi-bin/COnsIte/consite/). variants altering an amino acid, a splice site or a relevant transcription factor were considered as putative functional variants.
Intrafamilial segregation analysis
rare variants that altered an amino acid or that had been found in more than five patients were selected to assess if they were segregating or not with the PDB phenotype within families, to check if they could have been possible disease-causing mutations.
Case–control genetic association study for selected rare variants
Four rare genetic variants from 1p13 or 8q22 loci were selected for the case–control genetic association study. the selected rare variants were present among cases only and were meeting one of the following criteria: (1) the vari-ant caused an amino acid change predicted to be damag-ing by at least two in silico prediction tools among Poly-Phen, sIFt and Condel; (2) the variant had been identified in more than five cases in the discovery sample; or (3) the variant seemed to segregate with PDB in available family members. Figure 1 summarizes the steps that led to the selection of rare variants for the association study. seque-nom MassarraY snP Multiplex technology was used, as described above, for genotyping of these rare variants in 267 patients with PDB from the French-Canadian popula-tion and 295 healthy controls from the same population.
statistical analyses
respect of the Hardy–Weinberg equilibrium in the con-trol group was checked by a Chi-square test for conform-ity (data not shown). the snP rs10498635 within the RIN3 gene (locus 14q32), which did not meet this equilibrium (p = 0.005), was rejected from the genetic association study. this locus was not further investigated. according to the executable Quanto 1.2.4, the power of our case–control inde-pendent sample to detect a genetic association for a common variant with a significant relative risk ≥1.5 is 85 % in an addi-tive genetic model with a MaF of the risk allele of 0.25. the power of our sample to detect a genetic association for a rare variant with a significant relative risk ≥1.8 is 68 and 84 % in an additive genetic model with a MaF of the rare variant risk allele of 0.05 and 0.08, respectively. Comparison of MaF included Chi-square test for homogeneity with one degree of freedom, calculation of relative risk (rr) and calcula-tion of the confidence interval at 95 % (95 % CI) (Haldane 1956). the comparison of genotype frequencies included a
758 Hum Genet (2014) 133:755–768
1 3
Chi-square test for homogeneity with two degrees of free-dom or a Fisher’s exact test. the lathrop’s genotypic relative risks were estimated for variants with significant difference between allelic or genotypic distribution among cases and controls (p < 0.05) (lathrop and lalouel 1984). searches for genotype–phenotype associations, relying on Chi-squared or Fisher’s exact tests when appropriate for nominal values and t test for continuous variables, were also performed for snPs found to be associated with PDB in the literature. association studies were conducted on our entire cohort which consists of 36 patients carrying the SQSTM1/P392L mutation and 231 patients not carrying the mutation. In this paper, we focused on the results that were obtained in the subgroup “Paget with-out SQSTM1/P392L”. Conservative Bonferroni’s correction was applied for multiple testing.
Results
Genetic association study of snPs in novel loci reported to be associated with PDB
We replicated the allelic association of rs484959, rs499345 and rs10494112 (locus 1p13) and rs5742915 (locus 15q24)
with PDB in our population, in particular in the subgroup “Paget without the SQSTM1/P392L mutation” (table 1a). these allelic associations remained statistically significant after Bonferroni’s correction. neither allelic nor genotypic search for an association of rs4294134 (locus 7q33) met the significance threshold when considering the subgroup “Paget without SQSTM1/P392L mutation”. However, the same difference between MaF in cases and controls led to an allelic association of marginal significance (uncorrected p = 0.04) when including carrier of the SQSTM1/P392L mutation for the statistical analyses. although no allelic association was found for the rs2458413 (locus 8q22), a genotypic association was identified with PDB, in particu-lar for the heterozygous genotype AG versus both homozy-gous genotypes AA and GG, (uncorrected p = 6.9 × 10−4, rr = 1.73 [1.22–2.45]) (table 1b). searches for genotype–phenotype associations provided marginal associations in snPs of chromosomes 1 and 8 (supplementary table 3), whereas no association was found for the rs4294134 and rs5742915 (data not shown). the only genotype–phenotype associations which remained significant after Bonferroni’s correction was an association with male sex of the AA or AC genotypes of rs499345 (uncorrected p = 3.3 × 10−5, rr = 3.05 [1.78–5.24]) and of AG or GG genotypes of
≥5 carriersof the variant suffering from
PDB?
Intra-familial segregation study
Amino acid modification?
In silicoprediction for
functional effect?No NoNo
Carrier’s family is genetically informative?
Deleteriousmodification in ≥ 2 in silico prediction
tools
No
No
Yes
Segregation of the variant with
PDB? ≥1 control carrying the
variant?
Yes
Yes
Association study
No
Yes Yes
No
Yes
Yes
Random selection of a few variants in each category
Start
Identification of genetic variants
Sequencing of 10 candidate genes in the discovery sample
MAF≤ 5%or
Absent from EntrezSNP
No
Yes
Rare variants
Stop
Stop
Stop
Stop
Stop
End
Stop No
Fig. 1 steps towards the selection of rare variants for the association study
759Hum Genet (2014) 133:755–768
1 3
Tabl
e 1
res
ults
of
indi
vidu
al g
enet
ic a
ssoc
iatio
n an
alys
es b
y sn
Ps o
f no
vel l
oci t
o be
ass
ocia
ted
with
Pag
et’s
dis
ease
of
bone
in th
e lit
erat
ure
(a) a
llelic
ass
ocia
tion
Chr
Gen
essn
Ps‘t
otal
Pag
et g
roup
’‘s
ubgr
oup
Page
t with
out S
QST
M1/
P39
2L m
utat
ion’
Ma
FU
ncor
rect
ed p
rr
[IC
95
%]
Ma
FU
ncor
rect
ed p
rr
[IC
95
%]
alle
le c
ases
alle
le c
ontr
ols
alle
le c
ases
alle
le c
ontr
ols
(N =
550
)(N
= 5
94)
(N =
480
)(N
= 5
94)
1C
SF1,
EP
S8L
3rs
4849
59 G
>a
0.31
0.42
0.00
01*
0.62
[0.4
9–0.
79]
0.31
0.42
6.93
x 1
0−5 *
0.60
[0.4
7–0.
77]
1C
SF1,
EP
S8L
3rs
4993
45 C
>a
0.41
0.33
0.00
48*
1.41
[1.1
1–1.
80]
0.42
0.33
0.00
19*
1.48
[1.1
5–1.
90]
1C
SF1,
EP
S8L
3rs
1049
4112
a>
G0.
290.
220.
0033
*1.
49[1
.14–
1.95
]0.
310.
220.
0007
*1.
61[1
.22–
2.11
]
7N
UP
205
rs42
9413
4 a
>G
0.87
0.82
0.04
1.41
[1.0
1–1.
96]
0.87
0.82
0.24
1.22
[0.8
7–1.
70]
8T
M7S
F4
rs24
5841
3 G
>a
0.60
0.57
0.46
1.09
[0.8
6–1.
39]
0.60
0.57
0.47
1.09
[0.8
6–1.
40]
15P
ML
, GO
LG
A6A
rs57
4291
5 t
>C
0.53
0.43
0.00
07*
1.50
[1.1
9–1.
90]
0.54
0.43
0.00
04*
1.55
[1.2
2–1.
98]
(b)
Gen
otyp
ic a
ssoc
iatio
n
Chr
Gen
essn
PsG
enot
ype
‘tot
al P
aget
gro
up’
‘sub
grou
p Pa
get w
ithou
t SQ
STM
1/P
392L
mut
atio
n’
Gen
otyp
esU
ncor
rect
ed p
rr
[95
% C
I]G
enot
ype
Unc
orre
cted
p
rr
[95
% C
I]
Cas
esC
ontr
ols
Cas
esC
ontr
ols
(N =
275
)(N
= 2
97)
(N =
240
)(N
= 2
97)
1C
SF1,
rs48
4959
aa
vs
GG
+a
G21
481.
4 ×
10−
4 *0.
49[0
.29–
0.85
]16
487.
9 x
10−
5 *0.
33[0
.18–
0.60
]
EP
S8L
3G
>a
GG
vs
aa
+a
G12
393
0.00
2*1.
77[1
.26–
2.49
]10
993
0.00
2*1.
82[1
.28–
2.59
]
aG
vs
aa
+G
G13
115
60.
690.
82[0
.59–
1.14
]11
515
60.
780.
83[0
.59–
1.17
]
1C
SF1,
rs49
9345
aa
vs
CC
+a
C37
290.
201.
44[0
.86–
2.41
]33
290.
181.
48[0
.87–
2.51
]
EP
S8L
3C
>a
CC
vs
aa
+a
C88
132
7.6
x 10
−4 *
0.59
[0.4
2–0.
83]
7213
22.
0 x
10−
4 *0.
54[0
.38–
0.77
]
aC
vs
aa
+C
C14
913
60.
002*
1.41
[1.0
1–1.
96]
134
136
5.7
x 10
−4 *
1.51
[1.0
7–2.
12]
1C
SF1,
rs10
4941
12G
G v
s a
a+
aG
2318
0.01
91.
41[0
.74–
2.67
]21
180.
014
1.48
[0.7
7–2.
85]
EP
S8L
3a
>G
aa
vs
GG
+a
G13
718
60.
005*
0.59
[0.4
3–0.
83]
113
186
7.3
x 10
−4 *
0.53
[0.3
8–0.
75]
aG
vs
aa
+G
G11
593
0.02
51.
57[1
.12–
2.22
]10
693
0.00
6*1.
73[1
.22–
2.47
]
7N
UP
205
rs42
9413
4 a
>G
GG
vs
aa
+a
G19
720
30.
081.
32[0
.91–
1.91
]17
020
30.
121.
29[0
.88–
1.90
]
aa
vs
GG
+a
G2
120.
050.
22[0
.05–
0.97
]2
120.
080.
25[0
.06–
1.12
]
aG
vs
aa
+G
G65
790.
220.
89[0
.61–
1.30
]57
790.
270.
90[0
.61–
1.34
]
8T
M7S
F4
rs24
5841
3a
a v
s G
G+
aG
7810
10.
330.
80[0
.56–
1.15
]69
101
0.37
0.81
[0.5
6–1.
17]
G>
aG
G v
s a
a+
aG
2757
0.00
5*0.
48[0
.29–
0.78
]24
570.
008*
0.48
[0.2
9–0.
81]
aG
vs
aa
+G
G16
113
82.
2 x
10−
4 *2.
65[1
.92–
3.66
]14
113
86.
9 x
10−
4 *1.
73[1
.22–
2.45
]
CC
vs
tt+
Ct
7953
1.8
x 10
−14
*1.
91[1
.29–
2.84
]73
536.
3 x
10−
5 *2.
10[1
.40–
3.16
]
760 Hum Genet (2014) 133:755–768
1 3
rs10494112 (uncorrected p = 0.002, rr = 2.22 [1.33–3.70]) (supplementary table 3).
selection of candidate genes
a total of 18 candidate genes have been selected among the four PDB-associated loci in our population: seven genes in the 1p13 locus (ALX3, AMPD2, CSF1, EPS8L3, GSTM3, GSTM4, and PSMA5), three genes in 7q33 locus (C7orf49, CNOT4 and NUP205), three genes on the 8q22 locus [CTHRC1, LRP12 and TM7SF4 (which encodes DC-staMP)], and five genes in the 15q24 locus (CD276, PML, GOLGA6A, CCDC33 and UBL7) met the selection criteria, based on biological functions, genetic mice models and tissue expression data. all these genes were expressed in bone-active cells or their precursors and/or those involved in the nF-κB pathway, in the proteasome pathway, in autophagy or apoptosis, in bone cell function and/or sur-vival, in cellular adhesion (e.g., integrins), in intercellular communication (cytokines), or in the pathogenesis of other bone metabolic diseases (supplementary table 4).
Distribution of identified genetic variants by minor allele frequency
a total of 283 genetic variants were identified in our dis-covery sample; 135 were rare variants and 148 were com-mon variants (supplementary table 5). among the rare genetic variants, 38 were new variants that were not listed in EntrezSNP database (table 2) and 97 were also rare variants but already referenced in EntrezSNP or in 1,000 genomes database (supplementary table 6). 72 of the 220 variants for which a MaF reference value was available in EntrezSNP had a MaF below 5 %. this means that about a third of variants identified met our criteria of a rare genetic variant. the reference MaF was inferior to 2 % for ~50 % of rare variants. the real proportion of variants with a MaF inferior to 5 % was certainly higher than 32 % (72/220) since it is very likely that the majority of the variants for which no reference MaF value was available in EntrezSNP has a MaF below 5 %.
Distribution of rare variants found in patients with PDB according to their locations
among the 135 identified rare variants, 126 were found in at least one PDB patient. 55 of these 126 rare variants were located in the 1p13 locus; 32 in the 7q33 locus, 10 vari-ants were located in the 8q22 locus, and 29 in the 15q24 locus. 30 of the 126 rare variants found in at least one PDB patient were located in an exon (Fig. 2). nine other vari-ants carried by at least one PDB patient were located in an untranslated gene region. the remaining rare variants were Ta
ble
1 c
ontin
ued
(b)
Gen
otyp
ic a
ssoc
iatio
n
Chr
Gen
essn
PsG
enot
ype
‘tot
al P
aget
gro
up’
‘sub
grou
p Pa
get w
ithou
t SQ
STM
1/P
392L
mut
atio
n’
Gen
otyp
esU
ncor
rect
ed p
rr
[95
% C
I]G
enot
ype
Unc
orre
cted
p
rr
[95
% C
I]
Cas
esC
ontr
ols
Cas
esC
ontr
ols
(N =
275
)(N
= 2
97)
(N =
240
)(N
= 2
97)
15P
ML
, GO
L-
GA
6Ars
5742
915
t>
Ct
t v
s C
C+
tC
6496
0.01
50.
66[0
.45–
0.95
]56
960.
030.
67[0
.45–
0.98
]
tC
vs
tt+
CC
124
146
0.42
0.89
[0.6
4–1.
23]
102
146
0.15
0.81
[0.5
7–1.
14]
MA
F m
inor
alle
le f
requ
ency
, RR
rel
ativ
e ri
sk
* t
hese
unc
orre
cted
p v
alue
s re
mai
ned
stat
istic
ally
sig
nific
ant a
fter
con
serv
ativ
e B
onfe
rron
i’s c
orre
ctio
n (t
hres
hold
of
p va
lue
afte
r co
rrec
tion
= 0
.008
)
761Hum Genet (2014) 133:755–768
1 3
Tabl
e 2
nov
el r
are
gene
tic v
aria
nts
iden
tified
in th
is p
roje
ct
IDl
ocat
ion/
vari
ant d
escr
iptio
nIn
sili
co f
unct
iona
l eff
ect p
redi
ctio
nsM
inor
alle
le
freq
uenc
y
eM
Bl
acc
essi
on
num
ber
Chr
: gen
omic
lo
catio
n*G
enic
loca
tion*
nuc
leot
ide
am
ino
acid
type
of
amin
o ac
id c
hang
ea
min
o ac
id P
oly-
Phe
n (P
) SI
FT
(s)
C
onde
l (C
)
tra
nscr
iptio
n fa
ctor
s G
ain
(G)
Los
s (L
)
splic
e si
tes
Bra
nch
poin
t (B
) Sp
lice
sit
e (S
)
Cas
esC
ontr
ols
HG
0053
341:
1106
0331
1 11
0,60
3,31
23′
Ut
rc.
*43_
*44d
el
CC
inst
t0.
022
0
HG
0053
131:
1106
0312
83′
Ut
rc.
*227
t>
Cl
: aP-
10.
022
0
HG
0053
351:
1106
0303
73′
Ut
rc.
*318
C>
GB
: Bro
ken
0.02
20
HG
0053
151:
1101
6862
3In
tron
c.27
3-15
9C>
GB
: Bro
ken
00.
13
HG
0053
161:
1101
6962
4In
tron
c.63
7+90
C>
tG
: c-r
el
B: B
roke
n0.
017
0
HG
0053
141:
1104
5359
45′
Ut
rc.
-52G
>C
s: n
ew0.
017
0
HG
0053
368:
1043
9033
2e
xon
c.45
0a>
tp.
arg
150s
err
adic
alP:
Ben
ign
s: t
oler
-at
edG
: C/e
BPβ
00.
13
HG
0053
378:
1043
9043
0e
xon
c.54
8C>
tp.
Pro1
83l
eur
adic
alP:
Prob
ably
dam
agin
g s:
tol
erat
ed0
0.13
HG
0053
171:
1103
0687
8U
pstr
eam
c.-2
30-2
29C
>t
G: P
Par
γ0.
032
0
HG
0053
181:
1103
0473
3In
tron
c.-2
4-33
8C>
tB
: Bro
ken
0.03
10
HG
0053
191:
1103
0094
0e
xon
c.71
8G>
ap.
Glu
240l
ysr
adic
alP:
Ben
ign
s: t
oler
-at
ed0.
016
0
HG
0053
4415
:743
7312
1In
tron
c.84
+16
46 a
>G
s: n
ew0.
019
0
HG
0053
4515
:743
7310
2In
tron
c.84
+16
65 a
>t
s: n
ew0.
019
0
HG
0053
4615
:743
7304
3e
xon
c.11
8 a
>G
p.Il
e 40
val
Con
serv
ativ
es:
tol
erat
ed0.
019
0
HG
0053
4715
:743
7300
6e
xon
c.15
5a>
Gp.
asp
52
Gly
rad
ical
s: t
oler
ated
0.01
90
HG
0053
4815
:743
7102
5e
xon
c.21
0 G
>a
p. G
ln70
Gln
syno
nym
ous
0.01
80
HG
0053
4915
:743
6655
2e
xon
c.13
62 C
>t
p.a
sn45
4 a
snsy
nony
mou
sG
: sox
-60.
032
0
HG
0053
431:
1102
8300
9-
110,
283,
016
5′ U
tr
c.-1
07_1
00
insG
GG
GC
GG
0.01
70
HG
0053
201:
1102
8246
5e
xon
c.11
4G>
ap.
thr
38t
hrsy
nony
mou
s0.
015
0
HG
0053
211:
1101
9593
0U
pstr
eam
c.-3
09-2
773C
>a
0.01
70
HG
0053
221:
1101
9867
8U
pstr
eam
c.-3
09-2
5G>
as:
new
0.02
0
HG
0053
231:
1101
9870
95′
Ut
rc.
-303
G>
C0.
038
0
HG
0053
241:
1101
9994
0In
tron
c.17
7+39
G>
CG
: nF-
κB
s: B
roke
n0.
018
0
HG
0053
251:
1101
9997
1In
tron
c.17
7+70
G>
Cs:
Bro
ken
s:
Bro
ken
0.01
80
HG
0053
261:
1101
9997
2In
tron
c.17
7+71
a>
Cs:
Bro
ken
0.01
80
HG
0053
271:
1102
0015
5In
tron
c.17
8-57
G>
as:
Bro
ken
0.01
60
HG
0053
281:
1102
0415
33′
Ut
rc.
*277
C>
tl
: nr
F-2
0.01
60
762 Hum Genet (2014) 133:755–768
1 3
Tabl
e 2
con
tinue
d
IDl
ocat
ion/
vari
ant d
escr
iptio
nIn
sili
co f
unct
iona
l eff
ect p
redi
ctio
nsM
inor
alle
le
freq
uenc
y
eM
Bl
acc
essi
on
num
ber
Chr
: gen
omic
lo
catio
n*G
enic
loca
tion*
nuc
leot
ide
am
ino
acid
type
of
amin
o ac
id c
hang
ea
min
o ac
id P
oly-
Phe
n (P
) SI
FT
(s)
C
onde
l (C
)
tra
nscr
iptio
n fa
ctor
s G
ain
(G)
Los
s (L
)
splic
e si
tes
Bra
nch
poin
t (B
) Sp
lice
sit
e (S
)
Cas
esC
ontr
ols
HG
4264
968:
1055
0353
3e
xon
c.19
48G
>a
p.v
al65
0Met
Con
serv
ativ
eP:
Ben
ign
s: t
oler
-at
eds:
new
0.02
40
HG
0053
417:
1352
4241
0-7
: 13
5242
412
Ups
trea
mc.
-26-
258-
26-
256
delC
at
0.06
90.
13
HG
0053
407
:135
2425
91U
pstr
eam
c.-2
6-76
G>
a0.
017
0
HG
0053
397:
1352
6979
2-7:
1352
6979
4In
tron
c.12
18+
37_1
218
+
39d
elG
aa
s : B
roke
n0.
059
0.33
HG
0053
387
:135
2761
36In
tron
c.14
74-6
2G>
aB
: n
ew s
:
Bro
ken
0.01
50
HG
0053
291:
1099
6904
1U
pstr
eam
c.-2
1-68
G>
t0.
017
0
HG
0053
301:
1099
6880
6In
tron
c.29
+11
8C>
Gs:
new
0.01
60
HG
0053
311:
1099
6850
4In
tron
c.29
+42
0a>
G0
0.13
HG
0053
328:
1053
5135
3U
pstr
eam
c.-5
0-70
1G>
as:
Bro
ken
0.02
60
HG
0053
338:
1053
6158
4e
xon
c.80
4G>
ap.
Pro2
68Pr
osy
nony
mou
s0.
016
0
HG
0053
4215
:747
4165
4e
xon
c.75
5G>
ap.
arg
252H
isC
onse
rvat
ive
G: s
p10.
016
0
* G
enom
ic a
nd g
ene
loca
tions
acc
ordi
ng to
1,0
00 g
enom
es d
atab
ase
(firs
t tra
nscr
ipt)
763Hum Genet (2014) 133:755–768
1 3
identified n an intron or in the near gene region. numerous of these non-coding rare variants were predicted to have functional effects through splicing changes and/or tran-scription factors (such as nF-κB, nrF-2, sox-6, PParγ)-binding sites modifications that should be highly relevant to PDB pathogenesis (table 2 and supplementary table 6). 60 % of rare variants found in at least one PDB patient were located in functionally relevant gene regions.
In coding regions, two of the four exonic rare variants identified in 1p13 locus were missense variants. Both of them are located in EPS8L3 gene. the variant p.Met35Ile (rs17598321) is located in the PtB domain of the ePs8l3 protein. It leads to a conservative change predicted to be benign by PolyPhen and tolerated by sIFt even though this methionine is well conserved through evolution. the variant p.Glu240lys (HG005319) leads to a radical change also pre-dicted to be benign and tolerated. Four missense rare vari-ants were found in the 7q33 locus. the variant p.ala7Gly (rs17480616) in CnOt4 protein leads to a conservative change predicted to be benign and tolerated. two missense variants were found in position 373 of the nUP205 protein, p.Met373thr (rs58392569), a radical but benign and toler-ated change, and p.Met373Ile (rs61459701) which is a con-servative change. In addition, a radical change p.Gln1586His (rs140215067) of the nUP205 protein was predicted to be tolerated and benign. two of the five exonic rare variants identified in 8q22 locus were missense variants. the lrP12 variant p.val650Met (HG426496) modified an amino acid conserved through evolution. It is predicted to be benign and
tolerated. the tM7sF4 variant p.leu397Phe (rs62620995) is predicted to be deleterious by PolyPhen, but tolerated by sIFt. leucine in position 397 of tM7sF4 protein is con-served through evolution (Fig. 3). Five non-synonymous amino acid changes were found in the 15q24 locus. In the GOlGa6a protein, p.Ile40val (HG005346) was a conserva-tive and tolerated change whereas p.asp52Gly (HG005347) was a radical but tolerated change. the p.arg252His (HG005342) of the UBl7 protein was a conservative change. the p.val211Met (rs62641691) change in CD276 protein was conservative, but it was predicted to be prob-ably damaging according to Polyphen but tolerated by sIFt. the radical change p.arg717Cys (ePs_15_74336849) of the PMl protein was predicted to be tolerated and benign (sup-plementary table 6).
Genetic association study of rare variants
Four rare variants were selected for the genetic associa-tion study. When considering the subgroup “Paget with-out SQSTM1/P392l mutation”, marginal allelic asso-ciations were observed between PDB and rare variant c.372+259a>G (rs35500845), located in CTHRC1 gene and p.leu397Phe (rs6262099), located in DC-staMP pro-tein (encoded by TM7SF4 gene) (table 3a). the MaF of the c.372+259a>G (rs35500845) variant was lower in patients not carrier of a SQSTM1/P392L mutation than in controls (7.3 vs. 10.9 %, p = 0.046, rr = 0.65 [0.42–1.00]). How-ever, the distribution of genotypes between patients carrying the SQSTM1/P392L mutation and those not carrying this mutation was significantly different (p = 0.008). although not statistically significant, the MaF tended to be higher among patients with the SQSTM1 mutation than among not mutated patients (15 vs. 7 %, p = 0.056, rr = 2.29 [1.00–5.25]). Interestingly, this variant was more frequent in our sample composed of healthy individuals from the French-Canadian population (10.9 %) than in the reference sample (4.9 %) from EntrezSNP database. the global distribution of the p.leu397Phe (rs6262099) variant genotypes differed between patients and controls (p = 0.044; TT vs. CC+CT). the presence of at least one major allele (C) in the genotype suggested a protective effect against PDB (table 3b). the T allele frequency was twice as high in patients versus con-trols. the low frequency of this allele in our sample prob-ably explains why these differences did not meet the thresh-old for statistical significance.
Discussion
In this study, we replicated the allelic association of rs484959, rs499345, rs10494112 (locus 1p13) and rs5742915 (locus 15q24) with PDB in our population.
Fig. 2 Distribution of identified rare variants in patients with PDB according to their locations. UTR untranslated region
764 Hum Genet (2014) 133:755–768
1 3
In addition, a genotypic association was found for the rs2458413 (locus 8q22) for the heterozygous genotype AG versus both homozygous genotypes AA and GG (uncor-rected p = 6.9 × 10−4). Genotype–phenotype associations which remained significant after Bonferroni’s correction were: an association with male sex of the AA or AC geno-types of rs499345 (uncorrected p = 3.3 × 10−5) and of AG or GG genotypes of rs10494112 (uncorrected p = 0.002).
among the 18 candidate genes sequenced in this study, we identified 283 genetic variants. Of these, 135 met our preset criteria for a rare variant. among rare variants identi-fied in at least one PDB patient, 55 were identified in 1p13 locus, 32 in 7q33, 10 in 8q22 and 29 in 15q24 locus. the identification of as many rare variants was not surprising since half of the variants identified in the enCODe pro-ject had a MaF below 5 % (Gorlov et al. 2008). 30 rare variants were located in an exon. numerous non-coding rare variants were predicted to have functional effects through splicing changes and/or transcription factors (such as nF-κB, nrF-2, sox-6, PParγ)-binding sites modifica-tions that should be highly relevant to PDB pathogenesis and require further functional analyses. For example, the nrF-2 signaling which is associated with aberrant produc-tion of oxidative response genes (Xing et al. 2012), was recently reported to be involved when a SQSTM1/S349T mutation is present (Wright et al. 2013). Interestingly, we identified two rare variants which were predicted to modify nrF-2 binding sites in GSTM4 and PML genes, represent-ing a possible new pathway involved in PDB pathogenesis.
an allelic association of an intronic rare variant located in CTHRC1 (c.372+259a>G, rs35500845) was observed with PDB. the minor allele (G) of this rare variant was more frequent in controls (10.9 %) than in cases (7.3 %) (p = 0.046). However, MaF for this variant tended to be higher among carriers (15 %) than non-carriers (7 %) of the SQSTM1/P392L mutation (p = 0.056). since SQSTM1/P392L mutation has been shown to be a founder muta-tion supported by two different haplotypes in the French-Canadian population, we must be careful while interpreting this data; MaF of rare variants unrelated to the disease and its susceptibility might differ from the general population
and from the non-carriers of the SQSTM1/P392L mutation. Further study would be suitable to determine if this vari-ant can act as a modifier for the SQSTM1 gene. CTHRC1 stimulates the proliferation and differentiation of osteo-blast progenitors, by stimulating the expression of COL1A1 and Osteocalcin genes and of the gene coding for alkaline phosphatase (Kimura et al. 2008; leClair et al. 2007). It has also been shown that CTHRC1 is expressed in calcified atherosclerotic plaques, but not in healthy arteries (Pyagay et al. 2005). an increase in gene expression of CTHRC1 in PDB patients could explain the higher prevalence of blood vessel calcifications in these patients (laroche and Del-motte 2005; strickberger et al. 1987).
Genotypic association was also observed between PDB and an exonic rare variant (p.leu397Phe, rs62620995) located in the seventh transmembrane domain of the DC-staMP protein (Hartgers et al. 2000). this alteration is predicted to be probably damaging by PolyPhen. the geno-typic distribution of this variant differed between cases and controls (p = 0.044) when considering genotype frequen-cies in controls adjusted to the expected Hardy–Weinberg proportion. Cases were two times more likely to be hete-rozygous for this variant than controls (13/251: 5.6 % ver-sus 8/287: 2.7 %). TM7SF4−/− mice phenotype is compat-ible with a mild form of osteopetrosis. these mice do not possess multinucleated osteoclasts and they show uncou-pled bone remodeling (Yagi et al. 2005; Miyamoto 2011). since hypermultinucleation is an important characteristic of pagetic osteoclasts, the p.leu397Phe (rs62620995) vari-ant could contribute to the disease development by increas-ing the activity of the encoded protein DC-staMP, either by changing its expression or by disabling its mechanism of internalization (Mensah et al. 2010; singer et al. 2006).
this study led to the identification of several rare vari-ants in PDB patients. Until now, the role of rare genetic variants in PDB had never been studied. However, the con-tribution of rare variants in osteoporosis has been investi-gated. Four deleterious rare variants of the LRP5 gene were more frequent in a group of fracture-prone children (li et al. 2011). In addition, a rare variant of the WNK4 gene has been associated with low bone mineral density in the
Fig. 3 Conservation of the amino acid leucine (l) in position 397 of the DC-staMP protein, encoded by the TM7SF4 gene. the amino acid leucine (codon Ctt), which is modified into a phenylalanine (codon ttt) by the rare variant rs62620995, is framed in red in the
DC-staMP protein sequence from the COBalt database (http://www.ncbi.nlm.nih.gov/tools/cobalt/). leucine is highly conserved in evo-lution, suggesting that the rs62620995 variant can have an important functional effect on the function of the DC-staMP protein
765Hum Genet (2014) 133:755–768
1 3
Tabl
e 3
res
ults
of
indi
vidu
al g
enet
ic a
ssoc
iatio
n an
alys
es b
y ra
re g
enet
ic v
aria
nt
MA
F m
inor
alle
le f
requ
ency
, RR
rel
ativ
e ri
sk
* n
one
of th
ese
unco
rrec
ted
p va
lues
rem
aine
d st
atis
tical
ly s
igni
fican
t aft
er c
onse
rvat
ive
Bon
ferr
oni’s
cor
rect
ion
(thr
esho
ld o
f p
valu
e af
ter
corr
ectio
n =
0.0
13)
(a) a
llelic
ass
ocia
tion
Chr
Gen
e na
me
(var
iant
ID
)v
aria
nt lo
catio
n‘t
otal
Pag
et g
roup
’‘s
ubgr
oup
Page
t with
out S
QST
M1/
P39
2L m
utat
ion’
Ma
FU
ncor
rect
ed p
*r
r[9
5 %
CI]
Ma
FU
ncor
rect
ed p
*r
r[9
5 %
CI]
alle
le c
ases
alle
le c
ontr
ols
alle
le c
ases
alle
le c
ontr
ols
(N =
534
)(N
= 5
90)
(N =
462
)(N
= 5
90)
1A
MP
D2
(HG
0053
16)
c. 6
37 +
90
C>
t0.
000
0.00
0−
1.10
−0.
000
0.00
0−
1.28
−1
GST
M4
(rs6
5098
5)c.
457-
42 C
>t
0.04
90.
054
0.67
0.90
[0.5
3–1.
52]
0.04
80.
054
0.63
0.88
[0.5
0–1.
53]
8C
TH
RC
1 (r
s355
0084
5)c.
372
+ 2
59 a
>G
0.08
00.
109
0.10
0.72
[0.4
8–1.
08]
0.07
30.
109
0.04
60.
65[0
.42–
1.00
]
8T
M7S
F4(
rs62
6209
95)
c.11
89 C
>t
0.02
60.
014
0.13
1.91
[0.7
9–4.
59]
0.02
80.
014
0.09
2.06
[0.8
5–5.
01]
(b)
Gen
otyp
ic a
ssoc
iatio
n
Chr
Gen
e na
me
(var
iant
ID
)v
aria
nt
loca
tion
Gen
otyp
e‘t
otal
Pag
et g
roup
’‘s
ubgr
oup
Page
t with
out S
QST
M1/
P39
2L m
utat
ion’
Gen
otyp
esU
ncor
rect
ed p
*r
r[9
5 %
CI]
Gen
otyp
esU
ncor
rect
ed p
*r
r[9
5 %
CI]
Cas
esC
ontr
ols
Cas
esC
ontr
ols
(N =
267
)(N
= 2
95)
(N =
231
)(N
= 2
95)
1A
MP
D2
(HG
0053
16)
c. 6
37 +
90
tt
vs
CC
+C
t0
0−
1.10
−0
0−
1.10
−
C>
tC
C v
s t
t+
Ct
267
295
−0.
91−
231
295
−0.
78−
Ct
vs
CC
+t
t0
0−
1.10
−0
0−
1.10
−1
GST
M4
(rs6
5098
5)c.
457-
42C
>t
tt
vs
CC
+C
t0
00.
671.
10−
00
0.77
1.28
−
CC
vs
tt+
Ct
241
263
0.79
1.12
[0.6
5–1.
94]
209
263
0.74
1.15
[0.6
5–2.
04]
Ct
vs
CC
+t
t26
320.
880.
89[0
.52–
1.54
]22
320.
830.
87[0
.49–
1.54
]
8C
TH
RC
1 (r
s355
0084
5)c.
372+
259
GG
vs
aa
+a
G1
30.
320.
48[0
.049
–4.6
2]0
30.
100.
16−
a>
Ga
a v
s G
G+
aG
221
233
0.14
1.41
[0.9
1–2.
18]
194
233
0.08
1.53
[0.9
6–2.
43]
aG
vs
aa
+G
G40
580.
190.
74[0
.47–
1.15
]33
580.
140.
70[0
.44–
1.11
]
8T
M7S
F4
(rs6
2620
995)
c.11
89t
t v
s C
C+
Ct
00
0.06
1.10
−0
00.
041.
28−
C>
tC
C v
s t
t+
Ct
253
287
0.10
0.52
[0.2
1–1.
25]
218
287
0.08
0.48
[0.1
9–1.
17]
Ct
vs
CC
+t
t14
80.
101.
93[0
.80–
4.69
]13
80.
072.
09[0
.85–
5.13
]
766 Hum Genet (2014) 133:755–768
1 3
Portuguese population (Mendes et al. 2011). Few studies have attempted to establish an optimal method to study rare variants functional effects. Previously published studies had selected candidate genes playing a role in the pathophysi-ology of the disease or containing a common variant asso-ciated with a disease (Cohen et al. 2004; lusis and Paju-kanta 2008; nejentsev et al. 2009; Hershberger et al. 2008, 2010; Iascone et al. 2012). since the causal genetic variant responsible for a genetic association may be millions base pairs away from the associated common variant, it is likely that this second approach will fail to identify causal variants (Kent et al. 2011). Ways to prioritize identified genetic vari-ants within a gene depending on their frequency and their putative impact on protein function according to in silico prediction tools have already been suggested. However, most of the time, variants carried by one or more healthy individuals are rejected which is inadequate to identify rare variants (Hershberger et al. 2008, 2010; Iascone et al. 2012).
In the present study, candidate genes were selected based on their position within a locus associated with PDB and their potential role in the pathophysiology of the dis-ease. the strategy of sequencing a small sample of individ-uals seems appropriate to identify rare variants since 135 genetic variants meeting our preset definition of a rare vari-ant were identified. nevertheless, this study has some lim-its. Dna used for the project was extracted from periph-eral mononuclear blood cells, preventing detection of rare somatic variants in bone tissue. We may also have missed some PDB-associated variants located in non-coding regions since we only searched for rare variants in exons, intron–exon junctions and promoter regions. Because rare variants are thought to be population-specific, the results of this association study are not generalizable to other popula-tions. In a future study, a greater number of rare variants could be genotyped, in order to perform clustering of sev-eral variants in a single statistical analysis (asimit and Zeg-gini 2009). Moreover, since a large number of genetic vari-ants met the preset definition of a rare variant, we suggest that a lower threshold of MaF (<3 %) should be used in future projects aimed at identifying rare variants. the MaF reference value obtained from a european sample should be used instead of the combined MaF reference value.
In conclusion, this project has led to the identification of 135 rare genetic variants. Marginal genetic associations have been identified between PDB and two rare genetic var-iants located in the 8q22 locus in the French-Canadian pop-ulation. allelic association was identified between PDB and an amino acid change p.leu397Phe (rs62620995) in the DC-staMP protein, involved in multinucleation of osteo-clasts. Genotypic association was identified between PDB and the intronic variant c.372+259 a>G (rs35500845) located in the CTHRC1 gene, involved in the regulation of bone remodeling via its action on osteoblastogenesis. the
involvement of TM7SF4 and CTHRC1 in the pathophysiol-ogy of PDB deserves to be further investigated.
Acknowledgments Mariejka Beauregard was supported by a sum-mer program for medicine student award from the Canadian Institute of Health research, followed by a scholarship of the Fonds de recherche du Québec-santé (FrQ-s) for the Master. Dr. Michou is supported by a career award from the FrQ-s. this study was funded by the Canadian Institute of Health research (Catalyst Grant: Bone Health), the Fonda-tion du CHUQ, the Canadian Foundation for Innovation, the FrQ-s, the laval University and the CHU de Québec research Centre.
Conflict of interest the authors declare that they have no conflict of interest.
Web resources
Database
Cobalt http://www.ncbi.nlm.nih.gov/tools/cobalt/entrezGene http://www.ncbi.nlm.nih.gov/geneentrezsnP http://www.ncbi.nlm.nih.gov/snpGenatlas http://www.genatlas.orgGeneCards http://www.genecards.orgGeneloc http://genecards.weizmann.ac.il/geneloc/
index.shtmlHomoloGene http://www.ncbi.nlm.nih.gov/homologeneOMIM http://www.ncbi.nlm.nih.gov/omimPfam http://pfam.sanger.ac.ukrefseq http://www.ncbi.nlm.nih.gov/refseqUniProtKB http://www.uniprot.org/help/uniprotkbWikiGenes http://www.wikigenes.org
In silico prediction tools
Condel http://bg.upf.edu/condel/homeConsite http://asp.ii.uib.no:8090/cgi-bin/COnsIte/consiteensembl64 http://useast.ensembl.org/index.htmlHuman splicing Finder http://www.umd.be/HsFMicroCosm targets http://www.ebi.ac.uk/enright-srv/
microcosm/htdocs/targets/v5PolyPhen http://genetics.bwh.harvard.edu/pph2Primer3 http://biotools.umassmed.edu/bioapps/
primer3_www.cgiQuanto 1.2.4 http://hydra.usc.edu/GxesIFt http://sift.jcvi.orgtFsearch http://www.cbrc.jp/research/db/tFsearCH.htm
References
albagha OM, visconti Mr, alonso n, langston al, Cundy t, Dargie r, Dunlop MG, Fraser WD, Hooper MJ, Isaia G, nicholson GC, del Pino Montes J, Gonzalez-sarmiento r, di stefano M, tenesa
767Hum Genet (2014) 133:755–768
1 3
a, Walsh JP, ralston sH (2010) Genome-wide association study identifies variants at CsF1, OPtn and tnFrsF11a as genetic risk factors for Paget’s disease of bone. nat Genet 42:520–524 (pii:ng.562)
albagha OM, Wani se, visconti Mr, alonso n, Goodman K, Brandi Ml, Cundy t, Chung PY, Dargie r, Devogelaer JP, Falchetti a, Fraser WD, Gennari l, Gianfrancesco F, Hooper MJ, van Hul W, Isaia G, nicholson GC, nuti r, Papapoulos s, Montes Jdel P, ratajczak t, rea sl, rendina D, Gonzalez-sarmiento r, Di stefano M, Ward lC, Walsh JP, ralston sH (2011) Genome-wide association identifies three new susceptibility loci for Paget’s dis-ease of bone. nat Genet 43:685–689 (pii:ng.845)
asimit J, Zeggini e (2009) testing for rare variant associations in complex diseases. Genome Med 1:24. doi:10.1186/gm238
Bodmer W, Bonilla C (2008) Common and rare variants in multifac-torial susceptibility to common diseases. nat Genet 40:695–701 (pii:ng.f.136)
Bodmer W, tomlinson I (2010) rare genetic variants and the risk of cancer. Curr Opin Genet Dev 20:262–267. doi:10.1016/j.gde.2010.04.016
Cirulli et, Goldstein DB (2010) Uncovering the roles of rare variants in common disease through whole-genome sequencing. nat rev Genet 11:415–425. doi:10.1038/nrg2779
Cohen JC, Kiss rs, Pertsemlidis a, Marcel Yl, McPherson r, Hobbs HH (2004) Multiple rare alleles contribute to low plasma lev-els of HDl cholesterol. science 305:869–872. doi:10.1126/science.1099870
Frazer Ka, Murray ss, schork nJ, topol eJ (2009) Human genetic variation and its contribution to complex traits. nat rev Genet 10:241–251. doi:10.1038/nrg2554
Gibson G (2011) rare and common variants: twenty arguments. nat rev Genet 13:135–145. doi:10.1038/nrg3118
Gorlov IP, Gorlova OY, sunyaev sr, spitz Mr, amos CI (2008) shifting paradigm of association studies: value of rare single-nucleotide polymorphisms. am J Hum Genet 82:100–112. doi:10.1016/j.ajhg.2007.09.006
Haldane JB (1956) the estimation and significance of the logarithm of a ratio of frequencies. ann Hum Genet 20:309–311
Hartgers FC, vissers Jl, looman MW, van Zoelen C, Huffine C, Figdor CG, adema GJ (2000) DC-staMP, a novel multimem-brane-spanning molecule preferentially expressed by den-dritic cells. eur J Immunol 30:3585–3590. doi:10.1002/1521-4141(200012)30:12<3585:aID-IMMU3585>3.0.CO;2-Y
Hershberger re, Parks sB, Kushner JD, li D, ludwigsen s, Jakobs P, nauman D, Burgess D, Partain J, litt M (2008) Cod-ing sequence mutations identified in MYH7, tnnt2, sCn5a, CsrP3, lBD3, and tCaP from 313 patients with familial or idiopathic dilated cardiomyopathy. Clin transl sci 1:21–26. doi:10.1111/j.1752-8062.2008.00017.x
Hershberger re, norton n, Morales a, li D, siegfried JD, Gonza-lez-Quintana J (2010) Coding sequence rare variants identi-fied in MYBPC3, MYH6, tPM1, tnnC1, and tnnI3 from 312 patients with familial or idiopathic dilated cardiomyopathy. Circ Cardiovasc Genet 3:155–161. doi:10.1161/CIrCGenetICs.109.912345
Iascone M, Ciccone r, Galletti l, Marchetti D, seddio F, lincesso ar, Pezzoli l, vetro a, Barachetti D, Boni l, Federici D, soto aM, Comas Jv, Ferrazzi P, Zuffardi O (2012) Identification of de novo mutations and rare variants in hypoplastic left heart syndrome. Clin Genet 81:542–554. doi:10.1111/j.1399-0004.2011.01674.x
Kent JW, Farook v, Göring HHH, Dyer tD, almasy l, Duggirala r, Blangero J (2011) Do rare variant genotypes predict common variant genotypes? In: BMC Proceedings, p s87
Kimura H, Kwan KM, Zhang Z, Deng JM, Darnay BG, Behringer rr, nakamura t, de Crombrugghe B, akiyama H (2008) Cthrc1
is a positive regulator of osteoblastic bone formation. Plos One 3:e3174. doi:10.1371/journal.pone.0003174
laroche M, Delmotte a (2005) Increased arterial calcification in Paget’s disease of bone. Calcif tissue Int 77:129–133. doi:10.1007/s00223-005-0250-1
lathrop GM, lalouel JM (1984) easy calculations of lod scores and genetic risks on small computers. am J Hum Genet 36:460–465
leClair rJ, Durmus t, Wang Q, Pyagay P, terzic a, lindner v (2007) Cthrc1 is a novel inhibitor of transforming growth factor-beta signaling and neointimal lesion formation. Circ res 100:826–833. doi:10.1161/01.res.0000260806.99307.72
li WF, Hou sX, Ferec C, Chen JM (2011) Osteoporosis: Genet-ics. encyclopedia of life sciences (els). Wiley, Chichester. doi:10.1002/9780470015902.a0022925
lusis aJ, Pajukanta P (2008) a treasure trove for lipoprotein biology. nat Genet 40:129–130. doi:10.1038/ng0208-129
lyles KW, siris es, singer Fr, Meunier PJ (2001) a clinical approach to diagnosis and management of Paget’s disease of bone. J Bone Miner res 16:1379–1387. doi:10.1359/jbmr.2001.16.8.1379
Manolio ta, Collins Fs, Cox nJ, Goldstein DB, Hindorff la, Hunter DJ, McCarthy MI, ramos eM, Cardon lr, Chakravarti a, Cho JH, Guttmacher ae, Kong a, Kruglyak l, Mardis e, rotimi Cn, slatkin M, valle D, Whittemore as, Boehnke M, Clark aG, eichler ee, Gibson G, Haines Jl, Mackay tF, McCarroll sa, visscher PM (2009) Finding the missing heritability of complex diseases. nature 461:747–753. doi:10.1038/nature08494
Marian aJ (2012) Molecular genetic studies of complex phenotypes. transl res 159:64–79. doi:10.1016/j.trsl.2011.08.001
Mendes aI, Mascarenhas Mr, Matos s, sousa I, Ferreira J, Barbosa aP, Bicho M, Jordan P (2011) a WnK4 gene variant relates to osteo-porosis and not to hypertension in the Portuguese population. Mol Genet Metab 102:465–469. doi:10.1016/j.ymgme.2010.12.010
Mensah Ka, ritchlin Ct, schwarz eM (2010) ranKl induces het-erogeneous DC-staMP(lo) and DC-staMP(hi) osteoclast pre-cursors of which the DC-staMP(lo) precursors are the master fusogens. J Cell Physiol 223:76–83. doi:10.1002/jcp.22012
Miyamoto t (2011) regulators of osteoclast differentiation and cell-cell fusion. Keio J Med 60:101–105
Morales-Piga aa, rey-rey Js, Corres-Gonzalez J, Garcia-sagredo JM, lopez-abente G (1995) Frequency and characteristics of familial aggregation of Paget’s disease of bone. J Bone Miner res 10:663–670
nejentsev s, Walker n, riches D, egholm M, todd Ja (2009) rare variants of IFIH1, a gene implicated in antiviral responses, pro-tect against type 1 diabetes. science 324:387–389. doi:10.1126/science.1167728
Pyagay P, Heroult M, Wang Q, lehnert W, Belden J, liaw l, Friesel re, lindner v (2005) Collagen triple helix repeat containing 1, a novel secreted protein in injured and diseased arteries, inhib-its collagen expression and promotes cell migration. Circ res 96:261–268. doi:10.1161/01.res.0000154262.07264.12
roodman GD, Windle JJ (2005) Paget disease of bone. J Clin Invest 115:200–208
singer Fr, Mills BG, Gruber He, Windle JJ, roodman GD (2006) Ultrastructure of bone cells in Paget’s disease of bone. J Bone Miner res 21(suppl 2):P51–P54. doi:10.1359/jbmr.06s209
siris es, Ottman r, Flaster e, Kelsey Jl (1991) Familial aggregation of Paget’s disease of bone. J Bone Miner res 6:495–500
staden r (1996) the staden sequence analysis package. Mol Biotech-nol 5:233–241
strickberger sa, schulman sP, Hutchins GM (1987) association of Paget’s disease of bone with calcific aortic valve disease. am J Med 82:953–956
Wright t, rea sl, Goode a, Bennett aJ, ratajczak t, long Je, searle Ms, Goldring Ce, Park BK, Copple IM, layfield r (2013)
768 Hum Genet (2014) 133:755–768
1 3
the s349t mutation of sQstM1 links Keap1/nrf2 signalling to Paget’s disease of bone. Bone 52:699–706. doi:10.1016/j.bone.2012.10.023
Xing Y, niu t, Wang W, li J, li s, Janicki Js, ruiz s, Meyer CJ, Wang Xl, tang D, Zhao Y, Cui t (2012) triterpenoid dihy-dro-CDDO-trifluoroethyl amide protects against maladaptive car-diac remodeling and dysfunction in mice: a critical role of nrf2. Plos One 7:e44899. doi:10.1371/journal.pone.0044899
Yagi M, Miyamoto t, sawatani Y, Iwamoto K, Hosogane n, Fujita n, Morita K, ninomiya K, suzuki t, Miyamoto K, Oike Y, takeya
M, toyama Y, suda t (2005) DC-staMP is essential for cell-cell fusion in osteoclasts and foreign body giant cells. J exp Med 202:345–351. doi:10.1084/jem.20050645
Yan Jenny Chung P, van Hul W (2011) Paget’s disease of bone: evi-dence for complex pathogenetic interactions. semin arthritis rheum. (pii:s0049-0172(11)00199-5)
Zhang J (2000) rates of conservative and radical nonsynonymous nucleotide substitutions in mammalian nuclear genes. J Mol evol 50:56–68