Annual Report";
/* =============================================================================
Meta Data Configuration
==============================================================================
*/
$dc_description = "The John Curtin School of Medical Research - The Australian National University";
$dc_subject = "Australia, Canberra, Australian National University, John Curtin, School, Medical, Research, JCSMR, medicine, science, health, disease, cancer, hiv, diabetes, blood, genome, immunology, genetics, vaccine, drug, neuroscience, postgraduate, study, PhD, honours, student, nobel prize";
/* =============================================================================
Side bar Configuration
==============================================================================
*/
/* Is menu bar required? Set "1" for yes and "0" for no */
$sidebar = "0";
$sidebar_config = "/var/www/jcsmr/documents/_includes/cbis.config";
include ("/var/www/jcsmr/documents/_includes/header.inc") ;
?>
JCSMR Annual Report 2003 - John Curtin School of Medical Research, Australian National University
Back to Index
Professor Simon Easteal, Co Director |
|
Centre
for Bioinformation Science
The Centre for Bioinformation Science (CBiS) forms a bridge
between two areas of major strength at ANU, the mathematical and biological
sciences. CBiS brings together researchers with backgrounds in mathematics,
statistics and quantitative biology with the goal of developing a conceptual
architecture for an information-based, integrative approach to complex
biological systems.
Substantial interdisciplinary research activities continue, including
work on sequence analysis, genetic epidemiology, protein folding, comparative
genomics, molecular evolution, and statistical analysis of microarray
data. A feature of the work of the Centre has been the extent of our outside
collaborations. We now have well-developed collaborations with researchers
in the Centre for Medical Health Research (CMHR), the John Curtin School
of Medical Research (JCSMR), the National Centre for Epidemiology and
Population Health (NCEPH), the Research School of Biological Sciences
(RSBS), the Research School of Chemistry (RSC), the Research School of
Information Sciences (RSISE) and the Mathematical Sciences Institute (MSI),
the Research School of Physical Sciences and Engineering (RSPhysSE), some
of whom hold joint appointments with CBiS. CBiS has worked with the ANU
Supercomputing Facility (ANUSF) and the Australian Partnership for Advanced
Computing National Facility (APAC), and with development of teaching resources
in the School of Biochemistry and Molecular Biology (BaMBi) and the Department
of Mathematics.
These cross-campus collaborations provide a nucleus for
the development of further links. To encourage further collaborations
and improve the quality of bioinformatics practice on Campus, we undertook
two initiatives in 2003. The first was to support a half time position
to work in collaboration with staff in the Biomedical Resource Facility,
mainly concentrating on development of statistically rigorous methods
for analysis of microarray data. The second initiative was development
by CBiS of Graduate courses “Bioinformatics for Biologists”
organised by Gavin Huttley. We have also been heavily involved with outside
organizations and with the organization of outside activities.
A highlight of 2003 was the extremely successful BioInfoSummer
Symposium "Mathematical, Statistical and Computational Challenges
in Bioinformatics" organised by Hilary Booth (Chair), Yi Fang, Alex
Isaev and Sue Wilson.
|
Dr H Booth
Research during the year was undertaken on the following projects:
- An Efficient Z-score Algorithm for Assessing Sequence Alignments and
the POZITIV software (with S.Wilson, J. Maindonald, O. Nielsen and P. Maxwell)
An Iterative Approach to Determining the Length of the Longest Common Subsequence
of Two Strings (with S.Wilson, O.Nielsen, S. Macnamara (Hons student 2003),
M. Kahn and R. Brown)
- Modeling and bioinformatics studies of the human Kappa class Glutathione
Transferase predict a novel third Glutathione Transferase family with homology
to prokaryotic 2-hydroxychromene-2-carboxylate (HCCA) Isomerases (with A.
Robinson, G. Huttley, P. Board)
- Development of the statistics of the P-values of Identical K-words in
random sequences (with S.Wilson and P. Hall)
- Development of an EST pipeline for the fast throughput of coral genes
and the statistical analysis of their similarity to vertebrates and invertebrates
(with P.Maxwell, L. Grasso, D. Hayward, R. Saint, E. Ball and D. Miller
- Petri Net Models for Complex Biological Networks (with S.Wilson, M. Wakefield,
C. Burden, L. Staals, S. Roberts, R. Ball, S. MacNamara (PhD student, 2004)
and Raymond Chan (Honours Student, 2004)
Dr C Burden
In collaboration with S. Wilson and Y. Pittelkow, research has been carried
out on a statistical analysis of physically based adsorption models of oligonucleotide
microarrays. The hybridization of labelled target cRNA to short oligonucleotide
probes at the surface of microarray chips is believed to be driven by competing
processes of adsorption and desorption. Using data from the Affymetrix HG-U95A
Latin Square experiment, we have assessed the relative merits of various physical
models including equilibrium and non-equilibrium Langmuir models. This treatment
differs from existing analyses in the use of more relevant generalized linear
models. Particular attention is paid to questioning the validity of assuming
universality of parameters across genes, or across probes within a probeset.
We have also improved the accuracy of existing methods of inferring absolute
mRNA concentrations from probe intensity values using adsorption models.
Professor S Easteal
Research direction has continued to be aimed at understanding the basis of
a number of genetic disorders, and of the patterns of normal genetic variation
in human populations. A particular focus is the genetic basis of personality
variation, which is associated with a predisposition to common forms of mental
illness involving depression and anxiety. This work is done in collaboration
with the NH&MRC Centre for Mental Health Research.
Dr Y Fang
My work during 2003 may be considered under three main headings:
- Mathematical Modelling of Protein Folding. A more theoretical background
on interfaces between proteins and surrounding water has been developed,
concentrating on the contribution of surface tension to the free energy.
A paper was submitted.
- Geometric Statistics of Protein Structure. This is the initial stage
in preparing for prediction of protein structure using our novel mathematical
model. Many people are involved, Andrew Butterfield, Junmei Jing, Cath Lawrence,
and S. Wilson. The first statistics about distance between different parts
of a protein have been obtained and are being analysed. We will also launch
statistical studies about geometric characters defining globular proteins.
- Minimal Comprehensive DNA Sequence. With Zongzhu Lin of Kansas State
University, we have proved the existence of such sequences for all N-codons,
of length 4N/2 +N –1, for N odd, N≥1, For N even,
the ideal minimal length is (4N+4N/2-1)/2+N-1, but
we prove it is unreachable. Development of an algorithm for finding such
sequences is the next research goal. Due to the potential commercial value,
we are preparing to apply for a patent.
Dr G Huttley
A software toolkit aimed at comparative genomics has now largely been completed.
The toolkit is the basis for papers currently submitted, and ongoing research
projects. The outstanding questions currently being assessed are an examination
of sex specific mutation processes in humans; the influence of the DNA modification
5-methylcytosine on the pattern of divergence between coding and non-coding
DNA sequences; characterising the factors influencing sequence variation at
the breast and ovarian cancer susceptibility locus BRCA1 in the mammals; a
maximum-likelihood based phylogenetic footprinting approach; and, statistical
techniques for estimating functional dependence between sequence residues
using comparative sequence data.
Dr A Isaev
Bioinformatics research continued on models of dependent evolution of DNA
and protein sequences. Implementation is now in progress, and I helped resolve
many issues arising in the implementation process.
Dr J Maindonald
During 2003 I was involved in the following projects:
- Design of experiments, especially for cDNA microarray data.\
- Approaches to the analysis of cDNA microarray data; variance estimation
for the denominators of t and F-tests.
- Statistical software tools for use with cDNA microarray data, including
improved spatial plots.
- Use of the R system and associated packages for the analysis of microarray
data.
- Educational and training issues for laboratory scientists who work with
microarray data.
- Models and tests for Hardy-Weinberg disequilibrium, leading to the development
of an R package named “hwde”.
Estimation of between slide variance, for use in the denominator
of t-statistics and F-statistics, is a major issue for the statistical analysis
of cDNA microarray data. On the one hand, estimates that are based on individual
gene variation are, when the number of slides per treatment is small, highly
affected by sampling variation. On the other hand, systematic changes in variance
with average intensity, with print tip and with print tip order complicate the
attempt to combine variance information from multiple genes.
Dr R Turakulov
Analysis of genome scale SNP genotyping data.
Genetic strategies for understanding genotype-phenotype relationships in human
disease depend on assumptions about the architecture of variation in the human
genome. We have analysed patterns of SNP variation at >5,000 SNPs for samples
of African-American, Asian and Caucasian populations available from The Single
Nucleotide Consortia Ltd (TSCL) [http://snp.cshl.org/].
From this analysis two major conclusions were found: 1. Evaluation of the number
of SNPs required to reliably assign individuals to the population from which
they are derived. Our results show that less than 100 loci are required to identify,
with high confidence, the population source of individuals in these samples.
2. Identification of loci with unusually high levels of variation among populations.
A list of these loci has been generated. They are candidates for the action
of diversifying natural selection and they are highly informative indicators
of the population affinity of individuals.
Dr M Wakefield
The major research focus continued to be Comparative Genomics, with particular
emphasis on using divergent mammals to understand genes of medical importance.
The adaptive evolution of the breast cancer gene BRCA1 and the
evolutionary forces that have driven rapid change in this gene throughout the
mammalian radiation are the main focus of current research efforts. This year
saw the initiation of experiments to determine if differences in marsupial and
monotreme lactation and X chromosome inactivation have been the driving force
for waves of adaptive changes in BRCA1.
Through involvement in the ARC Centre for Kangaroo Genomics initial
work on the curration of genome mapping data and development of analysis methods
has proceeded in 2003, resulting in the development of a new maximum likelihood
based phylogenetic footprinting program for the identification of conserved
segments of the genome that are indicative of control regions.
Professor S Wilson
Research continued in statistical genomics and bioinformatics with emphasis
on the following projects:
- In the visualisation of gene expression data, development of further methods
for finding genes on microarrays that have similar patterns of up/down regulation
for samples, joint with Y. Pittelkow
- Improvements in the statistical approach to analysis of adsorption models
for oligonucleotide arrays, joint with C. Burden & Y. Pittelkow
- Exploration of various methods for finding the longest common subsequence
based on using an iterative approach, joint with H. Booth, R. Brown &
M. Kahn (ANU Supercomputer Facility)
- Collaboration on analysis of mapping data for interaction between loci for
diabetes onset in NOD mice, joint with C. Goodnow & A. Listen (JCSMR)
- Method found for detecting epigenetic parent of origin effects in candidate
genes, joint with J. Wicks (QIMR)
- Based on permutation tests, an empirical method (and code) has been developed
for evaluation of genetic and environmental interactions in complex disease,
joint with J. Maindonald, J. Cavanaugh & N. O' O'Callaghan (The Canberra
Hospital)
- Exploration of the relationship between a particular genotype, joint pain
and serum ferritin in hemochromatosis, joint with J. Cavanaugh & E. Walker
(The Canberra Hospital)
- Development of objective methods for determining appropriate reporting intervals
for analyte concentrations, joint with P. Hickman (The Canberra Hospital)
- Statistical modelling of non-bonding minimal atomic distances in globular
proteins, joint with Y. Fang, C. Lawrence & W. Kaplan (Garvan Institute)
- Examining novel methods for finding those genes underpinning the risk of
development of complex human disease/disorders, and understanding genetic
variation, with S Easteal and H Munroe
BioInfoSummer 1-5 December, 2003
"Mathematical, Statistical and Computational Challenges in Bioinformatics"
was the theme of the extremely successful Australian Mathematical Sciences
Institute (AMSI) Summer Symposium in Bioinformatics, hosted by the Australian
National University's Centre for Bioinformation Science, December 1-5. Sponsors
also included Cray, Ceanet, the Australian Partnership for Advanced Computing
and the Australian National University's National Institute for Bioscience
(NIB). The days were organised into Themes, allowing attendees who were unable
to attend the whole week to select those areas of particular interest. The
Themes were: Introduction to Molecular Biology; Sequences & Data; Evolutionary
Models & Genetics; Protein Structure & Function and Microarrays &
Experimental Design, the last theme being held jointly with the IBS Australasian
Region's biennial meeting. Each day started with Educational Lectures followed
by a mix of Keynote and Specialist talks. Most days had parallel sessions
in the afternoon, allowing those students who had registered for the Graduate
Course Award in Bioinformatics to do the Educational Computer Lab session
while others attended Specialist talks.
The keynote speakers were
John Mattick (Institute for Molecular Bioscience, Brisbane)
Terry Speed (Walter & Eliza Hall Institute, Melbourne)
Simon Easteal (ANU)
Warren Kaplan (Garvan Institute, Sydney)
Mark Ragan (IMB, Brisbane)
Gordon Smyth (WEHI, Melbourne)
Sue Wilson (ANU)
A poster session was held on Tuesday evening, and student scholarships were
awarded to 25 successful applicants. The ceremony was followed by a well-attended
public lecture by Jenny Marshall-Graves (ANU) on "Unravelling the Kangaroo
Genome". NIB prizes were awarded to Ann Kwan (best poster), Penny Bennett
and Natalie Thorne (best student talk, shared) and Jim Stankovich and Antonio
Reverter (best talk by a researcher within ten years of receipt of PhD, shared).