project title: developing mathematical modelling ... · centre for advanced imaging ... developing...
TRANSCRIPT
Project title:
Primary Supervisor: Further info:
Project title:
Primary Supervisor:
Developing Mathematical Modelling Techniques to Determine Protein Structure in solution
Associate Professor Jeffrey Harmer
Associate Professor Jeffrey Harmer Principal Research Fellow Centre for Advanced Imaging [email protected].
Neuroprotective and anti-inflammatory effects of complement inhibitors in traumatic brain injury model
Prof. David Reutens and Dr. Min Chen
Further info: Prof. David Reutens [email protected]
Dr. Min Chen [email protected]
Project title:
Primary Supervisor:
Further info:
Background field removal in ultra-low field MRI using machine learning
Professor David Reutens and Associate Professor Viktor Vegh
Associate Professor Viktor Vegh Principal Research Fellow Centre for Advanced Imaging [email protected]
Project title: Characterising the interactions between membrane-active peptides (MAPs) and lipid bilayers for rational design of MAP-based therapeutics
Primary Supervisor:
Dr Yanni Chin
Further info: Dr Yanni Chin Centre for Advanced Imaging [email protected]
2020 Winter Research Project Description
Project title: Developing Mathematical Modelling Techniques to Determine Protein Structure in solution
Project duration: 4 weeks Description: This research project will develop mathematical
methods and computer code to determine protein structure using a set of experimental distance constraints measured between pairs of spin labels attached to a protein with unknown structure. The experimental data is already available for a Non-Ribosomal Peptide Synthetase (NRPS), and our interest in this system is to understand how its structure-function relationship enables it to produce antibiotics – a very important goal for medical applications.
The project will generate a large set of trial NRPS structures in silicon which will be spin labelled in silicon and the set of distance distributions between the spin labels computed and stored. This data will then be fitted to the corresponding experimental distance distributions to define an optimal set of structures for the NRPS protein. To determine the optimal set regularisation techniques will be examined, for example,
A(p) = Σ||(yexp-Ysim.p)||2, (Eq. 1) where yexp are the experimental distance distributions, Ysim is a matrix of distance distributions generated from the trial protein structures, and p is the population probability of each conformation. As this is an under-determined problem, meaningful solutions can only be obtained by imposing additional regularisation constraints,
Q = A(p) + λR(p), with pi ≥ 0 (Eq. 2) where λ is a Lagrange multiplier that defines the relative weight of a regularisation function. The student will look at a number of algorithms, for example the maximum entropy algorithm where R(p) = −Σpilog(pi). The aim is to find a good fit that is statistically meaningful.
Expected outcomes and deliverables:
Scholars will gain skills in computer programming, data fitting and how this is used in a structural biology context.
Suitable for: This project is open to applications from students with a background in either chemistry, biochemistry, mathematics or physics. This project would be suitable for a person interested to learn about mathematical modelling and computer programming (we use MatLab).
Primary Supervisor:
Associate Professor Jeffrey Harmer
Further info: Associate Professor Jeffrey Harmer Principal Research Fellow Centre for Advanced Imaging [email protected].
A NRPS construct (top) and distance contrasts and fits used to construct the conformations (bottom)
20 40 60 80 100
r / Å
0
0.2
0.4
0.6
0.8
1
P(r)
/ a.u
.
20 40 60 80 100
r / Å
0
0.2
0.4
0.6
0.8
1
P(r)
/ a.u
.
20 40 60 80 100
r / Å
0
0.2
0.4
0.6
0.8
1
P(r)
/ a.u
.
20 40 60 80 100
r / Å
0
0.2
0.4
0.6
0.8
1
P(r)
/ a.u
.
20 40 60 80 100
r / Å
0
0.2
0.4
0.6
0.8
1
P(r)
/ a.u
.
2020 Winter Research Project Description
Project title: Neuroprotective and anti-inflammatory effects of complement inhibitors in traumatic brain injury model
Project duration: 4 weeks
Description: The complement system is a potent effector of innate immunity and a significant contributor to secondary tissue damage and to epileptogenesis following central nervous system injury. Our results have demonstrated that complement inhibitors, including IVIg, PMX-205 and CR2-Crry, significantly improved motor and cognitive outcomes in mouse model of controlled cortical impact. This project aims to examine neuroprotective and anti-inflammatory effects of complement inhibitors using histopathology. The brain tissues for this study are already available.
Histopathology. Serial brain sections will be stained as detailed below, followed by image acquisition and analysis using FIJI and/or Imaris software. Brain sections including the injured cortical region, hippocampus, and piriform and entorhinal cortices will be analyzed. Neuronal cell death and fibre sprouting. Neuronal cell death will be assessed with immunofluorescent staining for NeuN, MAP-2, Fluoro Jade B (Millipore) and cleaved caspase-3. Timm’s sulphide silver staining will also be used to quantitatively assess mossy fibre sprouting, a form of synaptic reorganisation in the dentate gyrus, using proportional area measurement. Gliosis, scarring and inflammation. Microglial/macrophage and astrocyte density will be examined by quantifying the number of nucleated (i.e. DAPI+) Iba-1+/CD11b+ and GFAP+ (glial fibrillary acidic protein) cells, respectively (immunofluorescence). Stains for M1 (CD 16/32) and M2 (CD 206) activated microglial phenotypes, neutrophils (Ly6G, neutrophil-specific clone 7/4) and T cells (CD3) will also be employed to further analyze the inflammatory infiltrate between conditions. Multi-colour staining for MAP-2, fibronectin and GFAP will be used to measure lesion core and peri-lesional volumes.
Expected outcomes and deliverables:
Scholars will gain skills in cryostat brain slice preparation, immunohistochemical staining, image acquisition and data analyse.
Suitable for: This project is open to applications from students with a background in biology, biochemistry.
This project would be suitable for a person interested to learn about the utilization and application of immunohistochemistry techniques.
Primary Supervisor:
Prof. David Reutens and Dr. Min Chen
Further info: Prof. David Reutens Dr. Min Chen [email protected] [email protected]
2020 Winter Research Project Description
Project title: Background field removal in ultra-low field MRI using machine learning
Project duration: 4-5 weeks
Description: In MRI the signal scales with the magnetic field strength of the scanner, and as such signal-to-noise ratio improvements can be gained through the use of higher and higher field strength instruments. Whilst these instruments provide excellent soft tissue contrast, they are not portable and are expensive to buy, maintain and operate. An alternative approach led by our group is the development of low cost, portable MRI instruments. This, however, means lower signals as we operate in the ultra-low field regime. To be able to recover some of the signal-to-noise ratio lost by going to lower fields, we have been working on methods which reduce noise.
Software gradiometry is a process by which noise can be reduced, if not eliminated, by the use of additional background field detectors and an appropriately matched software solution. Our interest has been to create a machine learning framework which takes input from many background field sensors and uses this information to remove noise captured by the sensor placed near the sample. In this project a number of background field sensors and machine learning approaches, such as convolutional neural networks and deep learning, will be applied for noise reduction.
Expected outcomes and deliverables:
The scholar will be exposed to real research in relation to ultra-low field MRI. It is expected that the person will design and implement either a convolutional neural network or a deep learning framework, optimise the network and assess its ability to remove background field / noise.
The project additionally involves data curation, analysis of outputs and the estimation of noise reduction ability of the network. At the end of the project, the student should be able to say how the various network features (e.g. number of layers, nodes within layers and layer arrangement) impact on noise reduction performance.
Suitable for: The scholar should have a physical sciences background (physics, maths, IT, engineering) with a background in artificial intelligence. The project is suitable for students who have in the least been exposed to convolutional neural networks and deep learning.
Primary Supervisor:
Professor David Reutens and Associate Professor Viktor Vegh
Further info: Associate Professor Viktor Vegh Principal Research Fellow Centre for Advanced Imaging [email protected]
The student will be provided with a workspace and software for implementing machine learning algorithms. Data for this project has already been collected, however the student may wish to run the ultra-low field MRI setup and collect new / additional data necessary to train and test the neural network.
2020 Winter Research Project Description
Project title: Characterising the interactions between membrane-active peptides (MAPs) and lipid bilayers for rational design of MAP-based therapeutics
Project duration: 4 weeks
Description: Background Peptides exhibit a wide range of biological activities. Their unique pharmacological profiles have made them attractive drug-leads for the development of novel therapeutics. Many biopeptides function by interacting with the cell membrane and that could involve disrupting it, passing through it, fusing with it or residing on it at the membrane interface. Peptides with these properties are referred to as membrane-active peptides (MAPs). They include antimicrobial peptides (AMPs) which kill cells by disrupting the structure of the membrane and some ion channel-modulating peptides (e.g. neurotoxins), which have a pharmacology that is significantly influenced by lipid binding. Understanding the mechanisms by which these peptides interact, partition or disrupt the cell membrane will guide the rational development of MAP-based therapeutics. The subject peptides in this study are Pn3a, a neurotoxin that is a potential analgesic drug-lead, and AA139 which has shown promising antimicrobial activities.
Aim To characterise the interactions between the MAPs and lipid bilayer using nanodiscs as the membrane-mimicking system
Approach NMR spectroscopy will be employed to characterise the interactions between the MAPs and lipid nanodiscs. The student will be involved in various aspects of protein expression and NMR characterisation.
Expected outcomes and deliverables:
Scholars will gain experience in recombinant protein expression and collecting and analysing data from NMR spectroscopy. Students may have an opportunity to generate publications from their contributions to this research project. Students will be expected to produce a report summarising the research at the end of their project.
Suitable for: This project is open to applications from year 3-4 students with a background in biochemistry and/or biophysics.
Primary Supervisor:
Dr Yanni Chin
Further info: Dr Yanni Chin Centre for Advanced Imaging [email protected]