| Project Title |
Oxford Nanopore reads correction |
| Project Code |
HPC_20_01084 |
| Principal Investigator |
Dr Frank Wellmer |
| Start Date |
2020-02-17 |
| End Date |
2022-10-25 |
| Abstract |
I am a PhD student in Frank Wellmer's Lab at the Department of Genetics.
Our area of research is flower development, mostly wet lab, but we recently started to use high throughput sequencing more consistently.
The aim of my project is to shed a light on the alternative splicing events that take place during the floral transition.
I performed various sequencing experiments with the MinION from Oxford Nanopore: the advantage of the technology is to obtain very long reads (in my case full-length transcripts), the downside is that the sequencer makes a lot more errors than an Illumina one.
To obviate this problem I want to correct the reads using a set of Illumina ones. This is a very computing intensive process (especially if you want high precision), 4 samples on our 6 core intel i7 desktop CPU would take at least a month, even when trading accuracy for speed.
I could really use some space on the cluster, for this project but also in the future; since it’s the first time that we face a problem like that I don’t know exactly how to proceed, I thought the best way was to explain my problem.
We are not a dry bioinformatics lab but a bit of computing power could really give a boost to my research.
|
|
| Project Title |
Electrophysiological markers of sematic processing during natural speech comprehension |
| Project Code |
HPC_20_01083 |
| Principal Investigator |
Assistant Professor Edmund Lalor |
| Start Date |
2016-09-01 |
| End Date |
2020-10-30 |
| Abstract |
This project investigates the neural correlates of semantic processing during the comprehension of naturalistic speech. Using computational language modelling and linear regression, we study how the brain uses semantic context to form predictions about upcoming words. |
|
| Project Title |
17/CDA/4760 - SoftEdge: Architectures & Algorithms for Software-Defined Edge Systems |
| Project Code |
HPC_20_01081 |
| Principal Investigator |
Prof George Iosifidis |
| Start Date |
2018-01-01 |
| End Date |
2021-12-31 |
| Abstract |
SoftEdge lies at the nexus of these game-changing ideas: (i) harnessing edge resources,
and (ii) deploying softwarised wireless systems enabled by the increasing programmability
and convergence of networks with IT infrastructure. Our goal is to develop a rigorous and
experimentally-validated framework for the design and control of multi-tier edge systems
that can be managed through an agile control plane (or, fabric), focusing on emerging
applications such as data analytics and rich multimedia services. |
|
| Project Title |
Changes in hippocampus volume in Alzheimer’s disease |
| Project Code |
HPC_20_01080 |
| Principal Investigator |
Dr Arun Bokde |
| Start Date |
2020-01-22 |
| End Date |
2020-10-23 |
| Abstract |
The development of Alzheimer’s disease (AD) is a slow process that takes 10-20 years before the symptoms appear – typically memory loss is the first initial symptom. A group of subjects at high risk of developing AD is Mild Cognitive Impairment (MCI) – subjects with mild memory impairments but otherwise are normal in activities of daily living. The earliest structure affected by AD is the hippocampus, and changes in hippocampus volume precede the appearance of memory impairments. The objective is to investigate the changes in hippocampus volume subfields in MCI groups that converted to AD as well as did not covert to AD.
|
|
| Project Title |
Quantum transport in 2D systems |
| Project Code |
HPC_20_01079 |
| Principal Investigator |
Dr Stephen Power |
| Start Date |
2020-01-14 |
| End Date |
2022-01-14 |
| Abstract |
2D materials offer a range of exciting electronic propertiess that can be exploited to developed new technologie.
Graphene in particular offers a range of advantages as a transport channel for currents, but other 2D materials are far better candidates for the generation and manipulation of spin and valley currents.
This project will use experimental-scale quantum transport simulations sto study the interplay of different materials, and to understand the experimental fingerprints behind fundamental physics at the nanoscale. |
|
| Project Title |
Machine learning on Graphene Nanoribbon |
| Project Code |
HPC_19_01078 |
| Principal Investigator |
Dr Stephen Power |
| Start Date |
2019-12-10 |
| End Date |
2020-12-10 |
| Abstract |
This project aims to investigate electronic behaviour in a new type of nanostructure, which combines
horizontal and vertical patterning of two-dimensional materials, and determine if this
system provides a platform to control exotic electron flavours
• to determine the exact role played by boundaries between different materials in these
systems and whether these boundaries can be tuned to control the flow of electrons with
certain properties
• to optimise the heterostructure geometries to perform complex operations using the
charge and flavour properties of electrons. |
|
| Project Title |
Quantum thermodynamics in non-Markovian settings using the TEMPO algorithm |
| Project Code |
HPC_19_01077 |
| Principal Investigator |
Prof John Goold |
| Start Date |
2018-09-01 |
| End Date |
2022-09-01 |
| Abstract |
The aim of this project is to apply the TEMPO method in order to understand how the laws of thermodynamics generalise to arbitrary quantum systems both at and away from equilibrium, and in both Markovian and non-Markovian settings. In particular, the algorithm has been updated from its original version in order to compute the quantum heat statistics by means of counting fields. So far the mean heat exchanged by the bath has been calculated for different temperatures, initial states of the system and values of the coupling strength between the bath and the system, in the case of a spin-boson model. Probability distributions of the heat exchange have also been computed. The next step will consist in implementing a second bath coupled to the system, which will allow calculations of heat currents and introduce the study of quantum thermal machines. |
|
| Project Title |
Which weeds shall inheret the Earth? Land use and Life History Strategy in the Anthropocene |
| Project Code |
HPC_19_01076 |
| Principal Investigator |
Professor Yvonne Buckley |
| Start Date |
2018-10-01 |
| End Date |
2022-10-31 |
| Abstract |
Human land use is the dominant global driver of ecological change. It causes community composition to shift, with species responding to disturbance in different ways. To conserve biodiversity and its services, we must understand and predict these differing responses. Detailed knowledge of most species’ demography and ecology is lacking, but can species’ general characteristics be used to predict population response?
Using data from over 9000 species in 6000 studies from global databases TRY and PREDICTS, I aim to assess the relationship between plant life form and population response to human land use. Life forms adapted to more disturbed environmental conditions are expected to be best able to persist in more intensive land uses.
This study will explore the utility of easily recorded traits present in large, open access databases for making meaningful predictions about the types of species likely to be affected by the main driver of ecological change, at a scale never previously investigated. |
|
| Project Title |
An Integrated Framework for Estimating the Number of Classes with Application for Species Estimation |
| Project Code |
HPC_19_01075 |
| Principal Investigator |
Prof Simon Wilson |
| Start Date |
2019-11-22 |
| End Date |
2020-02-29 |
| Abstract |
The current project is a simulation experiment needs to be completed within three months at most. It is part of my PhD thesis titled above and has the below abstract .
The two most common methods for estimating the number of distinct classes within a population are to either i) use sampling data directly with combinatorial arguments or to ii) extrapolate historical discovery data. However, in the former case such detailed sampling data is often unavailable, while the latter approach makes assumptions on the form of parametric curves used to fit the discovery data that is often lacking in theoretical justification. Instead we propose an integrated framework that allows the historical discovery data to be linked via proxy ‘activity data’ to infer a latent effort process. This process can then be used to form data on sampling records by way of constraints on how many samples were taken over time. Due to the nature of ‘data as constraints’ many inference techniques become infeasible, however, emulator-type approaches such as Approximate Bayesian Computation remain available. Such a process is demonstrated and explored through a number of simulated examples before we look at the consequence for recent global species estimates. |
|
| Project Title |
Optimal Control of Open Quantum Systems |
| Project Code |
HPC_19_01074 |
| Principal Investigator |
Dr Paul Eastham |
| Start Date |
2019-11-18 |
| End Date |
2020-11-01 |
| Abstract |
To find solutions to the problems posed by non-Markovian decoherence in quantum technologies. Objectives: (I)
Identify the appropriate models for promising quantum technologies whose performance is limited by non-Markovian
decoherence, and implement numerical simulations of these models. (ii) Develop machine-learning techniques for
identifying optimal control sequences for state preparation, coherence preservation, and the implementation of
quantum operations. (iii) Implement established optimization algorithms to identify optimal control sequences, and
compare their performance against machine-learning methods. |
|
