European Human Genetics Conference 2007 June 16 – 19, 2007 ...

Concurrent Sessions
C23. Literature-aided interpretation of microarray data: a
compendium study on muscle development and disease
P. A. C. ‚t Hoen 1 , R. Jelier 2 , E. Sterrenburg 1 , J. T. den Dunnen 1 , G. B. van Ommen
1 , J. A. Kors 2 , B. Mons 1,2 ;
1 Leiden University Medical Center, Center for Human and Clinical Genetics,
Leiden, The Netherlands, 2 Erasmus MC - University Medical Center Rotterdam,
Biosemantics Group, Department of Medical Informatics, Rotterdam, The Netherlands.
Lists of differentially expressed genes from microarray studies are difficult
to interpret. Comparative analysis of related microarray studies
may help to improve the understanding of the molecular and cellular
processes involved. We analyzed 102 microarray studies published in
the field of muscle development and disease. The overlap in the lists
of differentially expressed genes was generally small. Analysis at the
level of biological processes should be more rewarding, since different
genes may hint at the same process. Currently available methods rarely
performed well due to low sensitivity and their dependence on the
limited annotation of gene function and pathways in curated databases.
As an alternative, we developed a literature-based annotation system.
For all genes, species-independent concept profiles were derived, linking
all biological concepts mentioned together with the gene in Medline
abstracts in a weighted fashion. The concept profiles for all possible
gene pairs from two microarray studies were matched and evaluated
for significance of overlap. Our algorithm discovered biologically meaningful
associations between studies that had no genes in common. The
biological processes shared between studies were easily recognized.
For example, mouse and human studies on limb-girdle muscular dystrophy
2B were brought together by non-overlapping macrophage-expressed
genes. In vitro and in vivo studies into muscular atrophy were
connected by decreased mitochondrial activity, and found closely related
to mitochondrion-related myopathies. In conclusion, our approach
facilitates finding common biological denominators in microarray studies
in an unsupervised way, and without the need for raw data analysis
or annotations of biological functions in curated databases.
C24. Producing a reference human gene set
J. Loveland, J. A. Rajan, A. Frankish, E. A. Hart, J. G. R. Gilbert, L. Gordon,
E. Griffiths, R. Storey, S. Dyer, V. Curwen, S. Searle, L. Wilming, J. Harrow, T.
Hubbard;
Wellcome Trust Sanger Institute, Cambridge, United Kingdom.
The Havana group (www.sanger.ac.uk/HGP/havana/) produces manual
annotation of finished sequence from vertebrate genomes, concentrating
on finished human, mouse and zebrafish genomes. Manual
annotation is needed for the accurate determination of duplicated gene
clusters, pseudogenes and non-coding genes, polyA features, splice
variation and accurate nomenclature. Currently, annotation of genes
on the human genome is provided by multiple public resources. To
provide a standard uniform human gene set Havana is collaborating
with NCBI, UCSC and Ensembl to produce a consensus CDS (CCDS)
database containing a core set of mouse and human protein-coding
regions that are consistently annotated. The next release of the human
CCDS set will contain around 19000 coding transcripts compared to
13372 currently in mouse.
Havana also collaborates with Ensembl to improve the genebuild on
the finished genome and produce a merged gene set that is visible in
Ensembl. This merged set contains only full-length coding transcripts
predicted identically and independently by both groups and is coloured
gold in Ensembl. This enables Ensembl users to highlight errors within
the automated genebuild allowing Havana to correct the genes manually.
The resulting corrections get re-incorporated into the new genebuild.
The merged set consists of approximately 12,000 Ensembl and
Havana transcripts with identical exon coordinates and splice sites. It
differs to the CCDS approach in that UTR, as well as protein-coding
regions, must be the same. A new graphical interface called Zmap has
recently been implemented which enables multiple genomes to be annotated
simultaneously and thus should improve the consistency of
annotation.
C25. How healthy are children born after preimplantation genetic
diagnosis?
I. Liebaers, W. Verpoest, S. Desmyttere, M. De Rycke, C. Staessen, K. Sermon,
A. Michiels, K. Boelaert, J. Van der Elst, P. Devroey, M. Bonduelle;
Research centre for reproductive genetics, Vrije Universiteit Brussel – uzbrus-
21
sel, Brussels, Belgium.
Preimplantation genetic diagnosis (PGD) was introduced in the clinic
in 1990 as a very early form of prenatal diagnosis (PND) for couples at
high risk to transmit a genetic disease. Preimplantation genetic screening
(PGS), a variant of PGD was introduced later in order to improve
the selection of embryos for transfer in in vitro fertilisation (IVF).
The requests for PGD or PGS of infertile as well as of fertile couples
were evaluated and accepted for treatment whenever possible. As
soon as the PCR-based or FISH-based single cell assays were ready,
the IVF treatment was planned .A prospective data collection on pregnancies
and children was organised by giving questionnaires to the
couples on the day of transfer. Additional questionnaires were send to
the patients and their physicians (gyneco- logists, pediatricians) during
pregnancy, at delivery and later on. Children were examined at 2
months and 2 years of age whenever possible. Between 1993 and
2005, 2756 PGD and PGS cycles were performed for many different
indications including HLA-typing. After transfer of ‘fresh’ embryos, 567
children were born .Of these, 548 were liveborn, 19 were stillborn. And
9 children died neonatally. The perinatal death rate (5 from singleton
pregnancies and 23 from multiple pregnancies) of 4.9 %.should
be further investigated.. The mean term and birth weight in live born
singletons is respectively 38.8 weeks and 3268 grams. The major malformation
rate is 3.6%. Embryo biopsy does not increase the major
malformation rate when compared to children born after IVF/ICSI without
PGD.
C26. The use of proteomic methods to search for biomarkers in
maternal plasma from trisomy 21 pregnancies.
W. E. Heywood1 , T. E. Madgett2 , A. Wallington2 , J. Hogg1 , K. Mills1 , N. D.
Avent2 , L. S. Chitty1 ;
1 2 Institute of Child Health, London, United Kingdom, University of the West of
England, Bristol, United Kingdom.
Objective: To use proteomic methods to search for biomarkers for trisomy
21 (T21) in maternal plasma.
Methods: Surface Enhanced Laser Desorption Ionisation Mass Spectrometry
(SELDI) detects small mass proteins (