2008 NSW Bioinformatics Research Symposium

The 2008 NSW Bioinformatics Research Symposium is the continuation of a series of successful bioinformatics symposia coordinated by Sydney Bioinformatics over the last five years. Close to 150 people attended the 2007 Bioinformatics Research Symposium hosted by Sydney Bioinformatics and the University of Sydney Centre for Mathematical Biology.

The 2008 NSW Bioinformatics Research Symposium will be an opportunity to showcase:

• the latest bioinformatics research from universities and research institutes throughout NSW
• applications of bioinformatics research in critical areas of life science research

The Symposium provides a forum for both bioinformatics researchers and life science researchers to explore the latest in bioinformatics research, how these development apply in life science research, and establish cross-disciplinary research collaborations.

The 2008 NSW Bioinformatics Research Symposium is proudly sponsored by:

The NSW Bioinformatics Research Symposium will be held on:

Wednesday, November 26, 2008
from 8:50 to 5:10

in the

Footbridge Lecture Theatre
Camperdown Campus, University of Sydney

Lunch and refreshments will be provided.

There is no charge for this event but registration is compulsory as space is limited. Please complete the registration form and return to:

Dr Sonia Cattley
Level 1, Medical Foundation Building, K25
University of Sydney NSW 2008
Fax: 02 9036 3234
Email: sonia (at) sydneybioinformatics (dot) org

• Dr Jarno Tuimala - CSC Scientific Computing, Finland

Phylogeny of extant dinoflagellates and the evolution of nutrition

Dinoflagellates are parasitic, heterotrophic or photosynthetic flagellated protists, commonly regarded as micro-algae. They form red tides, and these mass-occurences are often characterized by the production of some very potent hepato- and neurotoxins. Currently, the phylogeny of dinoflagellates and the evolution of their nutrition is incompletely characterized. To further clarify these, phylogeny of dinoflagellates
(Dinophyta) was inferred using publicly available rRNA sequences. The results support the view that neither the "Pfiesteria-clade" nor the "Symbiodinium-clade" are monophyletic. This implies that parasitism might be both an ancestral and a derived state, whereas symbiosis is purely a derived state.

• Dr Gordon Smyth, Walter and Eliza Hall Institute, Melbourne

Statistical analysis of microarray expression data throws light on the immunological profile of early multiple sclerosis

This work is a collaboration with Trevor Kilpatrick, Vicky Perreau, Olga Skibina and coworkers at the Centre for Neuroscience of the University of Melbourne. Microarray transcription profiles were obtained of T cells from patients with a First Demyelinating Event, a condition which is a precursor to multiple sclerosis. Because of the relatively small sample sizes and patient-to-patient heterogeneity, "standard" microarray data analyses fail to detect differentially expressed genes, either between patients and controls or in response to treatment. On the other hand, a more developed analysis finds transciptional profiles for both disease and treatment as well as detecting patient subgroups. Gene set testing shows that FDE patients show a regulatory T cell signature. The analysis uses several recently developed statistical bioinformatics techniques for designed microarray experiments, including empirical quality weights, moderated random effects, and rotation gene set tests.

Download Dr Gordon Smyth's Presentation

• Dr Bret Church, School of Pharmacy, University of Sydney

A Knowledge-based Approach to Modelling Helical Proteins in a Layered Membrane

Integral membrane proteins are a class of protein with domains that traverse a membrane interface. Their location gives them fundamentally important roles for the cell, and hence makes them important drug targets, yet their environment is the very reason why experimental approaches used to determine three-dimensional structures may be thwarted. Only approximately 400 membrane protein structures exist in which the entire membrane is traversed, whereas the generalist structural database holds ~54,000 proteins. In this study, the approach of statistical potentials that have been successful with generalist structure modeling, are adapted in order to derive the 3-dimensional structures of such polytopic helical membrane proteins.

We report our helical membrane protein structure prediction method in which our set of scoring functions are applied to distinguish native-like polytopic helical membrane protein structures from non native-like structures. The scoring functions are trained and tested on 11 decoy sets. Each set consists of subdomains of 3-4 transmembrane helices, 66 to 145 residues in length, that were derived from known polytopic helical membrane protein structures. The novel aspects of the scoring function are the components designed for the membrane environment of the helices, and specifically the derived residue burial propensities and environment scoring. Although there is not extensive structural data for these proteins we have observed a regime free of bias to report the performance of the method. Each component of the scoring function performs successfully in identifying the corresponding features of native-like models. The weighting of the components were optimised on the decoy sets and the full scoring function with the weighted components was successfully tested on the available decoy sets. The program developed uses only the amino acid sequence of the target protein, but is also assisted by secondary structure and membrane topology predictions. The method used on five sequences gave predicted models with RMSDs of 5.0 to 12 Å from the native structures

Please contact the author directly regarding copies of this presentation.

---

• Dr Julianne Djordjevic, Westmead Millennium Institute

Delineating phosphatidylinositol specific phospholipase C-related virulence mechanisms of pathogenic fungi

Invasive fungal infections are an escalating global health problem and drug resistance is emerging. Understanding fungal virulence composites is therefore essential for developing new antifungal drugs. The model fungal pathogen, Cryptococcus neoformans, causes serious pulmonary and potentially fatal neurological infections in humans, affecting 30-40% of the 30 million HIV-infected patients in the developing world and causing 15-40% of their HIV-related deaths.

---

• Associate Professor Andrew Francis, School of Computing and Mathematics, University of Western Sydney

Exploiting genotypic data from tuberculosis outbreaks

Outbreaks of tuberculosis are often genotyped as a supplement to epidemiological methods. There are a number of genotyping technologies that are in use, but in some cases the technological advances have outpaced the development of analytical tools needed to interpret their output. In this talk I will survey some recent efforts to develop such analytical tools, and show some approaches that have been taken to analyse such data. For instance, we have used models of transmission together with evolution at the marker locus to estimate a number of epidemiological parameters using approximate Bayesian computation. These parameters include the basic reproductive number, the mutation rate of the marker, and the fitness cost of drug resistance. I will also discuss a method for determining whether one strain of the bacterium is spreading significantly faster than the background rate.

Copies of this presentation are available on request from Sydney Bioinformatics. Please contact Sebastien Gerega.

---

• Professor Rob Baxter, Kolling Institute

Biomarker discovery using protein-chip proteomics

The identification of protein markers of biological processes using conventional proteomic techniques is time-consuming and expensive when sample sets are large. A sensitive alternative with high sample throughput is afforded by surface-enhanced laser desorption-ionisation time-of-flight mass spectrometry (SELDI-TOF MS). In this technique, subsets of the proteome are captured according to their biochemical properties on protein chips with selective hydrophobic, ion-exchange, metal-affinity or ligand-affinity surfaces, and low molecular weight proteins (typically less than or equal to 25 kDa) are profiled as TOF mass spectra. Correlation between signal intensity and protein concentration is high for a given protein, and analytical precision is typically less than or equal to 25% CV. Despite the limitation of only interrogating subsets of the total proteome, this technique has been used successfully to identify putative disease biomarkers from cell culture, serum and tissue samples. Ovarian and breast cancer cell lines have revealed intracellular and secreted proteins that might act as putative biomarkers of key cancer-related events, such as high activity of PI 3-kinase signalling, or the induction of apoptosis by chemotherapeutic drugs. In collaborative studies we have also performed tissue and serum biomarker discovery from banked breast and pancreatic cancer specimens, using multivariate analysis to identify putative biomarker sets with high discriminatory power when tested independently. Candidate proteins can be directly eluted from protein chips or purified off-line for identification by LC/MS. Our studies to date suggest that that SELDI-TOF MS offers a highthroughput, analytically precise and reproducible approach to the identification of proteins that may be used to classify cancer and other patient groups. Supported by ARC, NHMRC, and the Cancer Institute NSW..

Please contact the author directly regarding copies of this presentation.

---

• Professor Marc Wilkins University of New South Wales

Title: Understanding the Dynamics of the Interactome

[Abstract: Interactome research has shown that proteins participate in a wide number and variety of interactions. Together, these form networks of great complexity. We have been constructing and visualizing these networks as scale-free graphs (1,2). A focus has been to use protein interaction networks as a scaffold onto which to project other biological data; this is useful both to understand the features of protein interaction networks as well as to understand the nuances of protein regulation. Data such as protein abundance, half-life, expression variance, subcellular localization and function can be analysed in this regard. We have also built views of the ‘complexome’, which is a higher-level organization of the cell than that of the interactome.

We have particular interests in what controls the dynamics of protein-protein interaction networks. We have defined three effects: sequence-based effects, abundance-based effects and conditional interaction effects (3) The latter is concerned with how post-translational modifications control protein-protein interactions and thus the delivery of particular functions inside the cell. We will discuss proteome-wide experiments that suggest methylation (4) as a major mechanism of controlling protein-protein interactions.]

---

• Associate Professor Pablo Moscato University of Newcastle

Towards robust classification methods for clinical bioinformatics

Robustness is increasingly being recognized as a central theme in Systems Biology. However, in the area of algorithm design and analysis, the robustness of the algorithms to perturbations of their inputs is seldom study or discussed. A main theme of our Centre is to develop algorithmic solutions which are aware of the problems that incorrect measurements can give to the final outcome. When these algorithms are used for clinical applications, a number of issues arise worth of consideration and mathematical modelling. We illustrate our talk with current progress in the application of our methods in neurodegenerative diseases, classification of tumours and prognosis evaluation of subtypes of cancer.

Please contact the author directly regarding copies of this presentation.

---

• Dr William Ritchie, Centenary Institute, University of Sydney

Dissecting microRNA function using multiple Bioinformatics approaches

microRNAs (miRNAs) are 20-22 nt small non-coding RNAs that reduce gene expression by decreasing mRNA levels and by inhibiting translation of "targeted" genes. These molecules were the subject of the 2006 Nobel Prize in Medicine and Physiology and play a major role in numerous functions such as development, disease, cell replication, cell repair, and apoptosis. Bioinformatics analysis has been instrumental in the discovery and functional characterization of these miRNA genes. However, because of their small size and the absence of a common primary sequence, the analysis of miRNAs necessitates the use of multiple bioinformatics approaches to achieve good results. These approaches encompass sequence conservation analysis, alternative expression studies and RNA structure analysis. In my talk, I will show how these techniques can be combined to 1) detect novel miRNA gense and 2) to characterize the function of known miRNAs. I will also briefly discuss the common pitfalls in bioinformatics analysis of miRNAs and suggest workarounds to be used by bioinformaticians working on miRNAs

Download Dr William Ritchie's Presentation

---

• Dr Kathy Belov University of Sydney

Exploring the genomes of marsupials and monotremes

[The recent sequencing of the genomes of the grey short-tailed opossum (Monodelphis domestica), the platypus (Ornithorhynchus anatinus) and the tammar wallaby (Macropus eugenii) has opened the door for gene discovery in our native wildlife. I will talk about the bioinformatic strategies that we have taken to characterize genes and gene families in these genomes. I will describe the discovery of novel antimicrobial peptides, involved in protecting immunologically naïve young. I will discuss the characterization of the platypus ‘venome’ using new generation sequencing and the search for an L to D-amino acid isomerase. The relevance of these discoveries to human immunity and health will also be discussed.

Please contact the author directly regarding copies of this presentation.