2009 Vienna - European Society of Human Genetics

Concurrent Sessions
c04.6
Duplication of the EFNB1 gene in familial hypertelorism:
imbalance in ephrin-b1 expression and abnormal phenotypes in
humans and mice
C. Babbs 1 , H. Stewart 2 , L. Williams 3 , L. Connell 3 , A. Goriely 1 , S. R. F. Twigg 1 , K.
Smith 3 , T. Lester 3 , A. O. M. Wilkie 1 ;
1 Weatherall Institute of Molecular Medicine, University of Oxford, Oxford, United
Kingdom, 2 Department of Clinical Genetics, Churchill Hospital, Oxford, United
Kingdom, 3 Genetics Laboratories, Churchill Hospital, Oxford, United Kingdom.
Familial Hypertelorism, characterised by widely spaced eyes, classically
shows autosomal dominant inheritance (Teebi type), but some
pedigrees are compatible with X-linkage. No pathogenic mechanism
has been described previously, but clinical similarity has been noted to
craniofrontonasal syndrome (CFNS), which is caused by mutations in
the X-linked EFNB1 gene.
Here we report a family in which females in three generations presented
with hypertelorism, but lacked either craniosynostosis or a
grooved nasal tip, excluding CFNS. DNA sequencing of EFNB1 was
normal, but MLPA indicated a duplication of all 5 exons of EFNB1,
which segregated with the hypertelorism. We characterised the duplication
breakpoint sequence, revealing a direct duplication of 937 kb
including EFNB1 and two flanking genes, PJA1 and STARD8. By use
of Pyrosequencing, we measured imbalance of EFNB1 expression in
the affected grandmother. After correction for skewed X-inactivation,
we show that the X chromosome bearing the duplicated EFNB1 genes
produces approximately twice as much EFNB1 transcript as the normal
X chromosome.
We propose that in the context of X-inactivation, the difference in expression
level of EFNB1 between the normal and duplicated X chromosomes
results in abnormal cell sorting during embryogenesis, leading
to hypertelorism. To support this hypothesis we provide evidence
from a mouse model carrying a hypomorphic Efnb1 allele, that abnormal
cell sorting occurs in the cranial region. Hence we propose
that X-linked cases resembling Teebi hypertelorism may have a similar
pathogenesis to CFNS, and that cellular mosaicism for different levels
of ephrin-b1 (as well as simple presence/absence) leads to craniofacial
abnormalities.
c05.1
combined analysis of 19 common validated type 2 diabetes
susceptibility gene variants show moderate discriminative value
and no evidence of gene-gene interaction
T. Sparso 1 , N. Grarup 1 , C. Andreasen 1 , A. Albrechtsen 2 , J. Holmkvist 1 , G. Andersen
1 , T. Jørgensen 3,4 , K. Borch-Johnsen 1,5 , A. Sandbæk 6 , T. Lauritzen 6 , S.
Madsbad 7 , T. Hansen 1,8 , O. Pedersen 1,7 ;
1 Hagedorn Research, Gentofte, Denmark, 2 Department of Biostatistics, University
of Copenhagen, Copenhagen, Denmark, 3 Research Centre for Prevention
and Health, Glostrup University Hospital, Glostrup, Denmark, 4 Faculty of Health
Science, University of Copenhagen, Denmark, 5 Faculty of Health Science, University
of Aarhus, Aarhus, Denmark, 6 Department of General Practice, Institute
of Public Health, Aarhus, Denmark, 7 Faculty of Health Science, University of
Copenhagen, Copenhagen, Denmark, 8 Faculty of Health Sciences, University
of Southern Denmark, Denmark.
Background: The list of validated type 2 diabetes susceptibility variants
has recently been expanded from three to 19. The identified variants
are common and have low penetrance in the general population. The
aim of this study was to examine for gene-gene interactions and investigate
the combined effect of the 19 variants by applying receiver operating
characteristics (ROC) to demonstrate the discriminatory value
between glucose-tolerant individuals and type 2 diabetes patients in a
cross-sectional population of Danes.
Methods: The 19 variants were genotyped in three study populations:
The population-based Inter99 study, the ADDITION study, and in additional
type 2 diabetic patients and glucose-tolerant individuals. The
case-control studies involved 4,093 type 2 diabetic patients and 5,302
glucose-tolerant individuals.
Results: Single variant analyses demonstrated allelic odds ratios (OR)
ranging from 1.04 (95%CI: 0.98,1.11) to 1.33 (95%CI: 1.22,1.45).
When combining the 19 variants subgroups with extreme risk profiles
showed 3-fold difference in risk of type 2 diabetes (lower 10% carriers
with 22 risk alleles, OR
2.93 (95%CI: 2.38,3.62,p=1.6×10 -25 ). We calculated the area under a
ROC curve to estimate the discrimination rate between glucose-toler-
ant individuals and type 2 diabetes patients based on the 19 variants.
We found an area under the ROC curve of 0.60. Two-way gene-gene
interaction showed few nominal interactions.
Conclusion: The 19 validated variants enables detection of subgroups
in substantial increased risk of type 2 diabetes, however the discrimination
between glucose tolerance and type 2 diabetes is still too inaccurate
to achieve clinical value.
c05.2
Joint re-analysis of twenty-nine correlated sNPs supports the
role of PcLO/Piccolo as a causal risk factor for major depressive
disorder
Z. Bochdanovits1 , A. van der Vaart2 , M. Verhage2 , A. Smit2 , E. de Geus2 , D.
Posthuma2 , D. Boomsma2 , B. Penninx1 , W. Hoogendijk1 , P. Heutink1 ;
1 2 VU Medical Center, Amsterdam, The Netherlands, Vrije Universiteit, Amsterdam,
The Netherlands.
The first genome-wide association study (GWAS) for major depressive
disorder (MDD) has implicated the pre-synaptic protein Piccolo,
but results from multiple replication cohorts remained inconclusive. We
propose a simple method for the joint (re-)analysis of multiple SNPs,
based on published summary data. Our approach is based on two
observations. Firstly, finemapping studies are focused, by design, on
a limited number of moderately to strongly correlated SNPs. All tested
SNPs are expected to reflect the true association of the unknown
causal variant proportional to their LD with it, in concordance with the
“Fundamental Theorem of the HapMap”. Secondly, given such correlated
SNP data it has been suggested before that a joint analysis of all
markers together is most powerful for detecting a true association. A
closer examination of the results reported in the GWAS study reveals
that the data indeed concur with the ”Theorem of the HapMap”. Based
on the above we re-analyzed the replication data using a novel joint
test of association and conclude and the results strongly favors Piccolo
to be a causal risk factor for major depression. This study was performed
within the framework of Top Institute Pharma project: number
T5-203.
c05.3
Unified framework for epistasis detection in (un)relateds
K. Van Steen1,2 , T. Cattaert1 , M. Calle3 ;
1 2 3 Montefiore Institute, Liège, Belgium, GIGA, Liège, Belgium, University of Vic,
Vic, Spain.
When searching for epistatic patterns parametric regression approaches
have severe limitations when there are too many independent
variables in relation to the number of observed outcome events.
Alternatively, the non-parametric Multifactor Dimensionality Reduction
method, MDR (Ritchie et al. 2001), can be applied. The common feature
of MDR and its extensions is that they are extremely computerintensive,
that best models are evaluated on the basis of cross-validation
(prediction accuracy measures) and permutations and that only
one such best model is proposed when looking at interactions of a
particular order.
We propose a novel unified multifactor dimensionality reduction strategy
for genetic interaction association analysis that can handle both
unrelated individuals and families of any structure, different outcome
types (e.g., categorical, continuous or survival type), easy covariate
handling or adjustment for lower order interactions or confounding
factors, all within the same framework. When applied to family data,
we obtain a less computationally intensive method than current MDR
adaptations to family-data and allow several clusters of markers to be
proposed as showing significant association with the outcome under
investigation. This better reflects locus heterogeneity and genetic heterogeneity,
usually present in complex diseases.
Our epistasis detection method is further evaluated and validated via a
simulation study, by computing type I error and power under a variety
of scenarios, and via application to a real-life data set.