Personal tools
\\\\\\\\\
Toolbox /////////
KAREN wiki
KIWI ADVANCED RESEARCH AND EDUCATION NETWORK
Fundamental, Mathmatical, Information and Computer Sciences
From KAREN wiki
Contents |
Fundamental, Mathematical, Information and Computer Sciences Case Studies
Carlo Laing
Case study: Complexity reduction in neural models.
Carlo Laing. Institute of Information and Mathematical Sciences Massey University.
c.r.laing@massey.ac.nz
Mathematical models of biological systems provide understanding and predictive power. However, as more information about a particular system is collected, its model becomes increasingly complicated. This makes the models better, because they are becoming more realistic and therefore potentially more useful, but in turn they are becoming harder to simulate and analyze. This research attempts to overcome this problem by using recently-developed, so-called “equation-free” techniques for the study of complex systems, concentrating on a variety of model neural networks. In particular the respiratory neural system is studied. This complex system regulates breathing by sending electrical signals to the diaphragm and other respiratory muscles, causing them to periodically contract and relax. Equation-free techniques allow one to automatically simplify a complex system such as the respiratory neural network by using judiciously designed computer simulations to capture the “essence” of the original complex system in terms of the correct variables. This provides insights and understanding that cannot be found in any other way. Detailed mathematical models of the respiratory system have been developed and studied but the exact mechanism(s) of rhythmic burst generation in the intact system are not yet well understood. While bifurcation theory and corresponding software have been successfully developed during the last few decades, the concepts involved are only really applicable to low-dimensional, deterministic systems. Realistic neuronal models are typically high dimensional, and sometimes stochastic. Equation free techniques allow one to computationally reduce such systems to deterministic low-dimensional systems and effectively apply bifurcation theory to the original complex system. This approach is used to study the respiratory neural system.
Research data is shared using compressed email attachments. Research work uses computer models, so data is generated and stored on computer.
Schematic of the respiratory neural network.
Elke Pahl
Case Study: From nano clusters to the bulk: Understanding melting from first principles.
Elke Pahl. Institute of Fundamental Sciences. Massey University E.Pahl@massey.ac.nz
Mercury is the only liquid metal at ambient conditions. This fact has been known since ancient times but the reasons for the exceptionally low melting temperature are still not well understood. Surprisingly, an accurate description of the melting of solids in general has not yet been solved in a satisfactory manner.
Recently research was able to obtain the melting points of the rare gases argon and neon with an unprecedented accuracy of less than 1 K deviation from the experimental values. The study used a cluster approach where the melting of finite atom clusters in dependence of their size was studied in a systematical manner by computer simulations. The results were then extrapolated to the infinite solid system. This is a generalist approach, our next goal is to address the fundamental question of mercury’s low melting point.
Simulations are computationally very expensive and rely heavily on Massey’s computer power (BestGrid and DoubleHelix).
Figure 1: Simulating the melting of the shown magic number rare gas nano cluster yields information about the melting of the corresponding infinite, solid system, if the cluster sizes studied are large enough.
Figure 2: The melting temperatures of the argon (Ar) and neon (Ne) cluster are extracted as the maxima of the heat capacity curves shown here. The insets display snapshots of the Ar147 cluster below and above the melting temperature.
Figure 3: Extrapolation of the cluster melting temperature for argon and neon to the infinite system and comparison with the experimental data.
Ian Bond
Case Study 1: Lowest mass star with orbiting planet. Dr Ian Bond.
Institute of Information and Mathematical Sciences. Massey University.
i.a.bond@massey.ac.nz
Software developed by Massey astronomer Dr Ian Bond is again at the centre of a planetary discovery with the detection of a planet 2-3 times the mass of the Earth orbiting a tiny “brown dwarf” star estimated to be only 6 per cent the mass of the Sun. The discovery is significant because it indicates that even the lowest mass stars can host planets and suggests that Earth mass planets are common throughout the Galaxy. Dr Bond says in the near future it may be possible to detect habitable Earth mass planets, with the new Kepler space mission that was launched by NASA in March, 2009.
The discovery is led by the Japan-New Zealand Microlensing Observation in Astrophysics (MOA) collaboration, which includes scientists at more than a dozen universities in New Zealand and Japan. Dr Bond is the New Zealand principal investigator for MOA. The aim is to discover planets orbiting distant stars using a technique known as gravitational microlensing which utilizes the star together any orbiting planets as naturally occurring lenses on the light of background stars.
A range of eResearch tools are used to gather, store and share data including:
• The World's largest telescope devoted to microlensing and supplied by Japan to the Mt John Observatory in New Zealand.
An 80 megapixel CCD camera supplied by Japan – one of the largest in the Southern Hemisphere.
• Image analysis software developed by Dr Bond for real-time processing of the large volumes of data generated by the telescope.
Researchers collaborate with one another and share information through a range of information networks such as instant messaging tools and software frameworks developed as part of the International Virtual Observatory initiative.
Case study: New Telescope to Study the First Generation Stars.
Dr Ian Bond. Institute of Information and Mathematical Sciences. Massey University.
i.a.bond@massey.ac.nz
Auckland Astronomers, including Massey University Astronomer Ian Bond, will use a new high technology telescope installed in Marlborough to study the first stars that formed in the Universe around 200 million years after the Big Bang. Collaborative research between Spain and New Zealand astronomers (Institute for Astrophysics of Andalucía (IAA-CSIC) and the Universities of Auckland, Massey and Canterbury and the Stardome Observatory) will attempt to detect these stars when the end their lives in spectacular “gamma ray burst” explosions resulting in the formation of a new black hole. If successful, these will be by far the most distant stars ever studied by astronomers and provide important new clues to the conditions in the early Universe. These events are first detected by NASA's orbiting satellites Fermi and Swift which immediately relay the position of the burst to ground-based telescopes. The new Marlborough telescope features advanced control software that will allow the telescope to lock onto new GRBs within seconds of detection by Fermi and Swift. The new telescope is robotic so once it is programmed for the night; it will be able to work unattended. It can also be operated directly via the Internet. eResearch tools used to support the research process (apart from the telescope) include new types of Web programming and instant messaging technologies to allow all the collaborating telescopes to interoperate across the Internet.
Inga Hunter
Case study: New Zealanders’ attitudes towards access to their Electronic Health Records.
Inga Hunter and Richard J Whiddett Department of Management, Tony Norris and Barry W McDonald, Institute of Information and Mathematical Sciences, and John A. Waldon, Centre for Māori Health Research and Development at Massey University.
This research project, funded by the Health Research Council, investigates public attitudes towards sharing confidential personal health information held in Electronic Health Records (EHRs). The project uses computer assisted telephone interviewing (CATI) to conduct a quantitative national survey of the attitudes of more than 4000 New Zealanders towards access to their personal health information using vignettes. Respondents are presented with vignettes which describe ways in which their health information might be used. The vignettes are in the form of: <person requesting information> would like to access your medical notes which contain <content of medical information> as <reason for request>. The information about you would contain <type of identification details>. After each vignette they are asked about their attitude to, and whether they would give consent for, such access. The final outcome of the project will be a specification of requirements for an e-consent model meeting the needs of most New Zealanders, thus enabling the potential benefits of electronically sharing confidential health information from EHRs to occur.
Preliminary results arising from the first 1828 respondents indicate that: • respondents were most willing to share their information for the purpose of providing care with doctors and nurses in a GP or hospital setting or with paramedics. Other health care provider groups were less popular. • However, removing their name and address so that information was anonymous, or identified only by record number, greatly increased the acceptability of sharing information for other purposes such as quality and financial audits and population based heath care.
Graph: proportion of respondents who agree to share their personal health information to provide care by ‘person requesting information’ (preliminary analysis of first 1828 respondents)
e-Research tools that enhanced or assisted with gathering, storing, analysing and sharing data and facilitated researcher collaboration included remote access to Massey university library databases, electronic collections and e-journals and video conferencing.
Justin O'Sullivan
Case study: Bioinformatic, statistical and molecular biological testing using novel e-research and molecular biological methods.
Justin O’Sullivan. Institute of Molecular Biosciences Massey University.
j.m.osullivan@massey.ac.nz
Imagine inserting one meter of fine rope into an empty potato-chip packet. As the rope coils and folds to fit the available space, various regions (loci) will come in contact -‘interact’- with each other. Like-wise, nuclei enclose single or multiple chromosomes comprised of DNA that is much longer than the diameter of the nuclear compartment. As a result of the limitations of nuclear space, interactions form within and between eukaryotic chromosomes. However, it is unknown if the majority of chromosomal interactions are accidental or functional (Figure 1) and therefore subject to selection.
We are using a novel combination of e-research and molecular biological methods to identify chromosomal interactions within Baker's yeast (Figure 2). The e-research tools that are central to our work were designed and written in house and run on standard computers. While the majority of our work runs on standard computers it can be constrained by memory limitations (on occasion requiring 64GB of RAM). We also use software that is available on-line and has been previously published.
Following the identification of interactions we perform bioinformatic, statistical and molecular biological tests to determine the significance of the interactions we have identified. Our novel approach to solving these problems is shifting the conceptual framework within which nuclei, chromosomes, and gene regulation are considered.
Mark Downey
Case Study: Development of Free Fat in Falling Film Evaporators.
Mark Downey and Tuoc Trinh. Institute of Food, Nutrition and Human Health Massey University.
m.c.downey@massey.ac.nz and t.trinh@massey.ac.nz
This research focuses on investigating damage to the milk fat globule membrane (MFGM) in milk powder manufacturing processes during evaporation in falling film evaporators used in the dairy industry. Damage to the MFGM was measured by a newly developed fatty acid test with complimentary measurements using particle size distribution data. All variables (flow rate and temperatures) were computer controlled through a PLC (Allen Bradley, CPU model SLC 5/03) and the variables recorded through a data acquisition programme (Intellution FIX DMACS 7.0). The temperatures were monitored using copper-constantan thermocouples or stainless steel resistance temperature detectors (RTDs). Pressures on the milk and steam sides were measured with inline sensors (Data Instruments, XPRO 25 or 50 psig) and flow rates were measured with an Endress-Hauser Picomag 11 PM16533 electromagnetic flow meter.
Richard Harris
Case study: Developed prototype planning tools for Telecom NZ.
Professor Richard Harris Institute of Information Sciences and Technology Massey University.
r.harris@massey.ac.nz
‘Telecom prototype planning tool’ research focuses on the resilient design of Next Generation Networks (NGNs) that carry a variety of network services, including voice services for commercial organisations. Next Generation Networks differ from the old Public Switched Telephone Network as they use Internet technology to deliver services such as voice, hence they share resources with Web-browsing and other data traffic on a single network infrastructure. The research process requires gathering a range of data associated with the entire voice service including: qualitative and quantitative service measures (e.g. user perceptions and physical voice traffic activity); summaries of network traffic; identification of network configurations, identification of the resilience and reliability of network switches and routers, individual and related services within the network and analysis of data in order to evaluate network strengths and points of network degradation. This information is used to develop service models that minimize the impact on customers of network system failures, boost traffic flows where required and improve and enhance voice network services using the NGN network. The research process includes developing a model that incorporates a variety of service parameters but also meaningfully measures service resiliency and design. The development of network models is enhanced through: • application and implementation of voice service technologies within the Next Generation Network • the development of telecommunication network planning software tools • gathering, storing and analysis of data electronically • collaborative sharing of data internationally using the KAREN network may be possible where it is not commercially sensitive.
Figure 1: Snapshot of initial planning tools developed at Massey
Figure 2: Sample output from a traffic study using the tool
Figure 3: Quality of Experience Model
Steve Flint
Case study: Bio-films in the food industry: problems and potential solutions.
Steve Flint. Institute of Food, Nutrition and Human Health, Massey University.
s.h.flint@massey.ac.nz
Contamination of food by spoilage and pathogenic micro-organisms costs the food industry millions of dollars annually. Much of this contamination may be attributed to the presence of bio-films in the processing plant. This research examines the properties of micro-organisms and the surfaces of processing equipment that influence the formation of bio-films. Of particular concern is the increased resistance of bio-films to cleaning and disinfection processes. The research identifies alternative means of controlling bio-film development.
Research findings indicate that a thorough understanding of the interactions of microorganisms with the surfaces of materials used in the construction of food processing plant is essential for the development of bio-film control regimes. The ad hoc approach to the design of cleaning systems has taken us as far as it can. Further improvements in the removal and control of bio-films can come only from strategies to target the binding of bacteria to the surfaces and to each other.
This research and its findings were enhanced by research data gathering, storage and sharing tools including nucleic acid data bases for microbial identification and web sites, in particular the Centre for Bio-film Engineering in Montana. Research collaboration was facilitated by the use of video conferencing facilities to link with collaborators at AUT and the University of Otago.
An illustration or diagram associated with the research – Thermophilic bacilli growing as a bio-film on a stainless steel surface.
Steve Stannard
Case study: Noninvasive Assessment of Hepatic Lipid Composition: Advancing Understanding and Management of Fatty Liver Disorders.
Steve Stannard et al. Institute of Food, Nutrition, and Human Health Massey University.
s.stannard@massey.ac.nz
Nonalcoholic fatty liver is frequently observed in obese, insulin resistant individuals, yet the factors that predict its development and progression to hepatic (liver) disease are poorly understood. This research identifies whether protein magnetic resonance spectroscopy might allow noninvasive assessment of hepatic lipid composition. Measurement of hepatic triglyceride concentration and composition using this assessment tool in healthy lean men, obese men with normal hepatic triglyceride concentration and obese men with hepatic steatosis indicates that: the proton magnetic resonance spectroscopy method provides rapid available assessment of lipid composition. Results are consistent with traditional invasive biopsy based methods. This readily available and noninvasive approach should promote further study into interactions between hepatic and whole-body lipid metabolism and help to elucidate the pathogenesis of disorders, such as type 2 diabetes, characterized by lipid accumulation within the liver. A range of eresearch tools contributed to and enhanced the findings of this research including: a protein magnetic resonance spectrometer to gather and analyze data, a combination of whole body (Q body) coil and a circular polarized surface coil for radio frequency signal transmission, a benchmark micro plate reader. Spectral data were post processed by magnetic resonance user interface software (jMRUI version 3.0). as well as the QUEST algorithm tool. Data was shared between researchers as large condensed emails.
- MRUI is a Graphical User Interface that allows MR spectroscopists to easily perform time-domain analysis of in vivo MR data. MRUI development (1998-2001) was supported by the European Communities project TMR/Networks ERB-FMRX-CT970160
Fig.4. (A) Hepatic triglyceride concentration (HTGC). (B) 1H-MRS measured hepatic lipid saturation index (SI) and (C) polyunsaturated index (PUI) in lean men (LEAN), obese men with normal hepatic lipid content (OB), and obese men with hepatic steatosis (OB_HS). Values are means _ standard error; n _ 10 for LEAN; n _ 8 for OB; n _ 9 for OBHS. *Significantly different versus LEAN (P _ 0.01). †Significantly different versus OB (P _ 0.01).
Wayne Patrick
Case Study: The evolutionary importance of being promiscuous (at the molecular level).
Wayne Patrick. Institute of Molecular Biosciences Massey University.
w.patrick@massey.ac.nz
Proteins perform essential roles in every biological process. Biochemistry textbooks typically describe them as being highly specific, accurate and very active at doing their jobs. These characteristics are usually associated with a lack of versatility. However, it is also apparent that proteins can evolve new functions rapidly – such as in the emergence of antibiotic resistance. An appealing solution to this conundrum is that proteins are promiscuous; that is, they have evolved to do one job extremely well, but by chance, they may be able to do other jobs (very poorly) too. These promiscuous activities could be the starting points for evolving new functions.
We are testing this model experimentally, by searching for promiscuous proteins and mimicking their evolution in the lab. To date, we have worked on cataloguing all of the promiscuous proteins in the model bacterium, Escherichia coli. We are discovering that protein promiscuity is much more common than previously appreciated. We are using the EcoCyc database of E. coli genes, proteins and metabolic pathways (www.ecocyc.org; Figure 1) to make sense of the experimental data.
One of our discoveries was that a protein called PurF could do the job of another protein called PRAI (in addition to its own job). This was surprising, as they are completely unrelated to each other. Nonetheless, we mimicked natural selection in the lab, and made a new PurF that was even better (by 30-fold) at doing its new job. We were able to explain the improvements in activity by careful consideration of the three-dimensional structure of the PurF protein (Figure 2). This was obtained from the Protein Data Bank (PDB; www.rcsb.org/pdb), which is the worldwide repository for biomolecular structures. This study showed that promiscuous proteins are able to change their functions rapidly, confirming their importance in molecular evolution.
Figure 1: Some of the metabolic pathways in E. coli, accessed through EcoCyc.
Figure 2: Mutations in the PurF protein that improve its PRAI activity. The protein was visualized using PyMol open source software (www.pymol.org).
Mike Tuohy
Glacier Meltdown Monitored from Space
Project Leader: Mike Tuohy (m.tuohy@massey.ac.nz)
The formation and expansion of the proglacial lake formed as Tasman Glacier retreats is well documented in a succession of satellite images. Over the past thirty six years, images of the region have been collected from various land resources satellites that orbit approximately 700km above the Earth’s surface.
These digital images require specialised software to convert the raw data into useful images from which accurate measurements can be made. The most important processing step is geo-referencing whereby the image is assigned real world co-ordinates like the New Zealand Map Grid. This allows images from different dates and different satellites to be overlaid and compared.
Here at Massey University the image processing is carried out in ENVI, ER Mapper or IDRISI Andes and the resulting images are incorporated into maps generated with ArcGIS.
A montage of satellite images showing the proglacial lake development. The lake area in 2008 is 489 hectares.
Ralf Toner
Massey computing power sheds light on amino acid interaction with surfaces
- Project Leader: Ralf Tonner
The understanding of interactions of peptides in our body with implants like artificial hips is crucial in determining factors that lead to occasional rejection of implants. Due to their non-toxicity and durability titanium based implants are often the material of choice. In the aqueous surrounding of the human body a titanium oxide layer is quickly formed which interacts with the peptides in the body. The building blocks of peptides themselves are amino acids.
Quantum chemical calculations can reveal the interactions of molecules and surface with unsurpassed accuracy. Since direct simulation of peptides interacting with a titanium dioxide surface is not feasible with our recent technology, studies have to rely on primer models to understand the underlying principles.
In the recent study we investigated the interaction of two amino acids – glycine and proline – with a titanium dioxide surface. For glycine we could resolve the discrepancy between the experimentally proposed adsorption mode for glycine (A) and the previously calculated mode (B) and found a new mode (C) in accordance with experimental data. In the first theoretical study of proline adsorption on titanium dioxide we found an adsorption mode which is in agreement with previous studies.
By using of the power of the recent computing clusters at Massey (BestGrid, DoubleHelix) we were able to propose properties of these bonding modes which can be probed experimentally. The approval of a new computing cluster (TripleHelix) will boost the available resources even further, enabling cutting edge research in computational chemistry.
Adsorption modes for glycine on titanium dioxide: A and B – previous findings, C – new finding in accordance with experimental results; D – proline on titanium dioxide.
