You will earn 6 research credits over 8 weeks, conducting faculty supervised, hands-on, directed study research projects with results that will culminate in the preparation of a research paper. You will complete a minimum of 240 hours on research in and out of the laboratory.
Faculty mentors will work closely with you to direct your continued growth and knowledge development in the chosen research topic discipline.
|DUBI RSLW 392S||International Independent Research in STEM Fields||6|
UCD Science Summer Projects 2018
Scanning digital information to identify probative value has been a timeconsuming task for forensic investigators. Innovative machine learning techniques and computer vision which are core branches of artificial intelligence, which show promise for digital forensic experts to carry out automated evidence classification. Accurate age estimation is crucial for a host of digital forensic applications and the potential has been proven for neural network based approaches to considerably outperform the capacity of human predictions. The best performing algorithms and services currently have a mean average error (MAE) rate of over 3.8 years. From our testing, this MAE gets significantly worse at the extremes of age ranges, i.e, <18 and >60. Of course, age prediction has a myriad of applications outside the field of digital forensics. Adult Entertainment venue access, purchase of adult goods, age targeted advertising and lately, recognition of children refugees from Syria with services such as how-old.net are just some of the multiple applications that can be achieved with multi-layered deep learning technologies. As part of the UCD Forensics and Security Research Group, this project is a collaboration with current researchers working to reliably and intelligently automate much of the digital forensic process. To date, we have curated large datasets of annotated human subjects to train and develop automated, deep-learning techniques for improving automated age estimation. We are focused on developing novel age estimation techniques using off-the-shelf visual processing hardware, such as the Intel Movidius Neural Compute Stick and cloud services, combined with our software defined age estimation approaches.
A common requirement in digital forensic investigations is to compare the contents of a seized device with a set of "known-illegal" files in order to identify pertinent evidence. For identical copies of known files, traditional hashing techniques suffice. However, traditional hashing fails where alterations have been made, either intentionally or through normal computer usage, and an alternative approach is required.
Approximate Matching (AM) algorithms have the capability of finding matching files even though alterations have been made (e.g. lengthening or shortening, embedding images into Office or PDF files, etc.). This project will perform a study of the types of scenarios where AM can be useful, and what types of changes it cannot detect. It will initially involve creating a dataset that includes files for a range of such scenarios. This will include files have been truncated, extended, combined, and edited in various ways. An initial set of files that can be adapted for this purpose will be provided. Following this, a number of experiments will be conducted to compare and contrast the capabilities of the three most prominent current AM techniques: mrsh-v2, ssdeep and sdhash. Implementations of each of these have been made available by their respective authors.
The outcome of the project will be to indicate where future research in this area should focus. This will be done through identifying scenarios where approximate matching is not suited to finding evidence where certain types of alterations have been made to files.
In psychology literature, the experience of agency describes a person’s innate sense of being in control of their own actions and, through this control, affecting the external world. At a very basic level this means that if we press a light switch and a light turns on we experience both control of the action and a responsibility for the outcome. In more complex situations the experience of agency can be less direct. This is particularly true when people are interacting with intelligent computer interfaces. Here the computer may use machine learning techniques to infer human intentions and then either assist in completing an action or even act independently of human action. The aim in this project will be to implement and conduct a series of experiments that use psychological techniques to measure variations in the human experience of agency when interacting with intelligent computer interfaces. If
the work is successful we will prepare a paper for submission to a relevant computer science conference, e.g. ACM CHI 2019
The number of cases requiring digital forensic analysis has significantly increased in recent years. Often, individual cases require the acquisition and analysis of digital evidence from a multitude of sources to gain any valuable insights. These sources can include computer equipment, mobile devices, file synchronisation servers, cloud servers, web service data, social media, wearables, navigation equipment, instant messaging, email, among others. The sheer volume of potentially evidence-rich data, alongside the complexity involved in accessing and acquiring the data from such a variety of sources leads to extended processing time per case. Digital forensic investigators typically find themselves spending 100-200 hours working each case. Current digital forensic analysis techniques are overly arduous and lack sufficient intelligent automation.
The goal of this project is to develop and train Deep Learning based artificial neural networks to enable the automated processing of digital forensic evidence, i.e., evidence categorisation and photo/video object identification, etc. The European Cybercrime Centre (EC3) often resort to social media in an attempt to crowd source insight into their cases, e.g., identifying a hotel from a hotel room photo, product/clothing identification, etc. Each step of the automated process must be verifiable and reproducible to hold up to cross examination in court, while ensuring the sensitive handling of privileged data. The successful creation of Deep Learning techniques for digital evidence processing will ensure that digital investigators have more time to investigate criminal activity, as their time can be used more effectively solving cases.
Dr. Aonghus Lawlor
Recommender systems often try to explain their recommendations to the users. The explanations can lead to greater trust between the user and the recommender system and can assist the users in making better decisions overall. The net result is an increased conversion from users to buyers and a better satisfaction rate among customers. We have built a framework to generate explanations for the recommendations in a dataset derived from Beer Advocate where users have provided ratings of various beer attributes (appearance, taste, aroma, etc) and also text reviews expressing their opinions about the beer. In this project we will explore techniques for personalising the explanations to the specific requirements of the user. We use a range of techniques, including natural language processing, opinion mining and sentiment analyses to build detailed feature-based product descriptions, which form the basis for our explanations. The goal of this project will be to design and build rich explanations for beer recommendations, using the Spark platform to scale the processing load.
EXPECTED - extend the existing explanation generation framework to scale to significantly larger datasets and to produce explanations with very low latency - design and build rich explanation structures which work with the specialised item classes in the Beer Advocate dataset
BACKGROUND - Muhammad, K., Lawlor, A., Rafter, R., & Smyth, B. (2015). Great Explanations: Opinionated Explanations for Recommendations. In Case-Based Reasoning Research and Development (pp. 244-258). Springer International Publishing
Software Requirements - python, (familiarity with pandas is a plus) or scala, text analytics (nltk), scikit-learn, (py)spark
Dr. Vivek Nallur
Emergence is a phenomenon that has been observed in many systems, physical, chemical, biological, socio-technical, etc. However, emergence is hard to pin down to a specific set of features/conditions. Hence, emergence detection is a hard problem, especially in complex adaptive systems. The DETECT framework proposes a decentralised mechanism to cooperatively detect emergence, without any previous knowledge of the domain and very little human intervention. However, it has only been shown to work on systems that are fairly homogeneous (in terms of the entities involved) in nature. For example, traffic jams
involving taxis, flocking of birds, spontaneous lane formation in pedestrians. For a truly useful framework, it is essential to know whether DETECT’s feedback-loop approach works even when the agents involved are heterogeneous in their abilities, interests and goals.
Experiment: The project will implement the DETECT framework and attempt to detect emergence in large-scale heterogeneous adaptive systems.
Pre-requisites: The student must have a working knowledge of statistics and be fluent in Python / Julia / Elixir.
 Decentralised Detection of Emergence in Complex Adaptive Systems – O’Toole et al.
Dr Breandán Kennedy
The goal of the UCD Ocular Pharmacology & Genetics group is to develop treatments for blindness based on enhanced understanding of the genes and drugs that modify visual function. In this project, the student will characterise novel CRISPR-Cas9 zebrafish models of inherited blindness and assess the potential of pharmacological interventions to overcome disease phenotypes. Techniques implemented include: zebrafish stock
maintenance, zebrafish larval genotyping (PCR, RFLP, SNP-PCR, DNA sequencing), analysis of retinal histology (light microscopy), analysis of visual function (optokinetic and visualmotor behavioural assays) and drug treatment of larvae.
1: Daly C, Shine L, Heffernan T, Deeti S, Reynolds AL, O'Connor JJ, Dillon ET, Duffy DJ, Kolch W, Cagney G, Kennedy BN. A Brain-Derived Neurotrophic Factor Mimetic Is Sufficient to Restore Cone Photoreceptor Visual Function in an Inherited Blindness Model. Sci Rep. 2017 Sep 12;7(1):11320.
2: Smith AJ, Carter SP, Kennedy BN. Genome editing: the breakthrough technology for inherited retinal disease? Expert Opin Biol Ther. 2017 Oct;17(10):1245-1254.
3: Merrigan SL, Kennedy BN. Vitamin D receptor agonists regulate ocular developmental angiogenesis and modulate expression of dre-miR-21 and VEGF. Br J Pharmacol. 2017 Aug;174(16):2636-2651.
4: Butler CT, Reynolds AL, Tosetto M, Dillon ET, Guiry PJ, Cagney G, O'Sullivan J, Kennedy BN. A Quininib Analogue and Cysteinyl Leukotriene Receptor Antagonist Inhibits Vascular
Endothelial Growth Factor (VEGF)-independent Angiogenesis and Exerts an Additive Antiangiogenic Response with Bevacizumab. J Biol Chem. 2017 Mar 3;292(9):3552-3567.
Dr Siobhán McClean
People with cystic fibrosis suffer from chronic infections throughout their lives, leading to a gradual but unrelenting decline in their lung function, which is ultimately the cause of death . The most difficult of these infections is caused by Burkholderia cepacia complex (Bcc), a highly antibiotic resistant group of bacteria that causes chronic opportunistic lung infections in people with cystic fibrosis. The cystic fibrosis lung is a low-oxygen environment. We have shown that Bcc evolves over time of infection in response to low oxygen and improves its ability to attach to lung cells , thereby improving its ability to colonise the lung. We have identified a group of stress proteins that are stimulated in the low oxygen environment of the cystic fibrosis lung, which are upregulated during chronic infection in cystic fibrosis patients. The aim of this project is to examine the role that these stress proteins play in chronic infection.
This project will involve comparing mutants that lack individual stress proteins with those that express the stress proteins. We will use a proteomics approach to examine the pathway that these stress proteins are involved in. We will also investigate the impact that low oxygen stress has on other proteins and functions in the bacterium. Investigating this process is important because if we can unlock the mechanism by which these bacteria evolve and enhance their ability to colonise, we can target the pathway and prevent chronic infection in people with cystic fibrosis.
Cullen, L. and S. McClean, Bacterial Adaptation during Chronic Respiratory Infections. Pathogens, 2015. 4(1): p. 66-89.
Cullen, L., A. O'Connor, P. Drevinek, K. Schaffer, and S. McClean, Sequential Burkholderia cenocepacia Isolates from Siblings with Cystic Fibrosis Show Increased Lung Cell Attachment. Am J Respir Crit Care Med, 2017. 195(6): p. 832-835.
Dr. John O’Connor, Associate Professor.
Acute hypoxia is associated with numerous central nervous system (CNS) disorders including stroke, Alzheimer’s disease and obstructive sleep apnea. In the CNS It results in a decrease in synaptic transmission which may be fully reversible upon re-oxygenation. Stabilization of hypoxia-inducible factor (HIF) by inhibition of prolyl hydroxylase domain (PHD) enzymes has been shown to regulate the cellular response to hypoxia and confer neuroprotection both in vivo and in vitro. Hypoxic preconditioning has become a novel therapeutic target to induce neuroprotection during hypoxic insults. However, the effects of repeated hypoxic insults or pharmacological PHD inhibition on synaptic signalling remain unresolved. In this project we will assess the effects of hypoxic exposure and PHD inhibition on synaptic transmission in the rat CA1 region. Field excitatory postsynaptic potentials (fEPSPs) will be elicited by stimulation of the Schaffer collatoral pathway. 30 min hypoxia will be applied to brain slices to inhibit synaptic transmission and reversed after 15 min with re-oxygenation. It is expected that 15-30 min of hypoxia will be sufficient to induce stabilization of HIF in hippocampal slices. Exposure to a second hypoxic insult after 60 min will be carried out in the pre condictioning experiments. Slices will be pre-treated with the PHD inhibitor, dimethyloxalyl glycine also to investigate if recovery is increased, reduced or unaltered. We hope these investigations into hypoxia and ‘pseudohypoxia’ preconditioning may improve our understanding of the processes that occur in many CNS disease processes.
Dr. Derek Costello
Background: Alzheimer’s disease (AD) is characterised by the age-associated accumulation of amyloid-β peptide (Aβ) within the brain, leading to the neuronal dysfunction and cognitive impairment. Aβ is a potent activator of microglial cells, inducing inflammatory mechanisms similar to those activated by certain bacterial infections. Interventions which target inflammatory mediators have proven successful in attenuating Aβ-induced changes in microglia and the associated disruption of hippocampal synaptic function. In addition, microglial activation promotes hyperexcitability of hippocampal neurons, which facilitates the cognitive decline and neurodegeneration characteristic of the disease.
Many patients with sporadic AD, as well as AD animal models, experience recurrent spontaneous seizures which are thought to precede cognitive impairment. Overexpression of Aβ promotes the development of epileptic activity, in turn accelerating cognitive decline. The role of inflammation in epileptogenesis, following disease, trauma and infection, has been highlighted in recent years, which can convey resistance to current anti-epileptic therapies. Enhanced production of inflammatory mediators following seizure further potentiates the activity. This is attributed to the over-activation of microglia and altered ability of astrocytes to maintain homeostatic function, thus lowering the threshold for neuronal activation.
Aims and Objectives: Our preliminary evidence indicates that certain bacterial agents, in combination with Aβ, differently impact the generation of epileptiform activity in hippocampal neurons. This study will build on this information, to analyse the inflammatory impact of HMGB1, a molecule produced in response to brain trauma, which acts in a similar manner to Aβ. We will investigate inflammatory changes in cultured microglial cells following exposure to these agents, and the potential impact that this may have on neuronal viability. In addition, we will examine the integrity of neurons that have been modified to produce and express different forms of Aβ, as an in vitro model in which to study Alzheimer’s disease pathology.
Experimental design: Microglial and neuronal cells will be cultured and maintained in vitro. Cells will be incubated in the presence or absence of HMGB1, a known mediator of cell damage. The resulting synergistic inflammatory response will be evaluated. Microglial cells will be assessed using PCR and Western immunoblot analysis to determine markers of inflammatory activation and signalling. Supernatants will be analysed by ELISA for evidence of soluble inflammatory mediators. Neurons will be assessed for viability and alterations in homeostasis
Techniques: In vitro electrophysiology, Western immunoblot analysis, ELISA, PCR Lab
Campylobacter jejuni is the leading cause of bacterial gastroenteritis in humans. This bacteria is found in large numbers in chickens where it causes no apparent disease and chicken meat in supermarkets can thus contain high numbers of these organisms which can then act as a reservoir of infection of humans. This bacteria displays significant genotypic variation and recent studies from our laboratory have revealed significant genetic and phenotypic variation between a small number of strains isolated from chicken meat. This project will aim to isolate more bacteria from chicken bought in irish supermarkets which can then be phenotypically and genotypically characterised. Techniques will include PCR, SDS Electrophoresis, Motility assays and a variety of bacterial culture techniques.
Dr Margaret Mc Gee
Tumour cells evolve from a primary benign cellular mass to cells exhibiting a phenotype of increased proliferation and eventual progression to a metastatic growth, where the tumour spreads to a distant site, or organ, that is distinct from the primary growth. Metastatic tumours are extremely difficult to treat and are often resistant to current chemotherapeutic agents. Thus, 90% of deaths due to cancer occur as a result of metastatic growth. Research in the Mc Gee Lab is focussed on understanding the molecular mechanism of metastasis with the view to developing new effective therapeutics for the prevention of metastasis.
Recent evidence highlights a critical role for cancer-derived exosomes during metastasis. Exosomes are small vesicles that are released from the primary cancer cells into the extracellular space. These extracellular vesicles, which contain DNA, RNA and protein from the cell of origin, are taken up by cells at another location within the body where they initiate a signalling cascade to promote metastasis. We have discovered part of the mechanism by which these vesicles are formed in cancer cells, and we have developed new drugs that could block vesicle formation, and therefore the process of metastasis, which we are testing in a variety of cancer types.
Research work in the lab involves growing a range of cancer cells in culture (glioblastoma, lung, breast and colon cancer) and isolating extracellular vesicles that are secreted from the cells. The amount and content of vesicles released will be determined following treatment with our new drugs. Multidisciplinary techniques commonly undertaken in the lab include molecular biology (DNA/RNA isolation, PCR), mammalian cell culture, Drug treatment, DNA electrophoresis, protein isolation, SDS-PAGE, Western blotting, Cell Viability and Cell Death assays and molecular imaging.
Associate Professor Carl Ng
We have identified a novel dwarf T-DNA mutant in B. distachyon, the new temperate grass model species. Preliminary experiments showed that mutant plants possess larger stomata and lower stomatal densities. In this project, we will conduct detailed phenotypic characterisation of stomatal size and densities using impressions obtained from leaves, and from isolated epidermis. We will conduct bioassays to determine maximum stomatal pore areas and water loss characteristics following leaf detachment. In addition, we will perform complementation analyses to determine the underlying genetic mechanisms contributing to the novel phenotype. This is a lab-based project involving phenotypic and molecular biology techniques.
Dr. Gavin Stewart
Bacterial fermentation processes in the human colon produce large quantities of short chain fatty acids (SCFAs), such as butyrate. These SCFAs are a vital source of energy, particularly for colonic epithelial cells, and are reabsorbed across cell membranes by a number of different mechanisms. One absorption pathway involves a group of epithelial proteins known
as monocarboxylate transporters (MCTs), which are capable of transporting SCFAs across cell membranes. Two key examples of these transporters are MCT1 and SMCT1. These transporters have been shown to play crucial roles in human colonic health and are known to be affected in inflammatory bowel disease, cancer and may even contribute to obesity. Our recent laboratory studies have identified that the expression of MCT1 and SMCT1 varies in different segments along the human colon, as well as in different disease states.
Using various biological research techniques, this project will investigate the effect of SCFAs on the expression, abundance and localization MCT1 and SMCT1 in various human intestinal cell lines (e.g. Caco-2, HT-29, FHC and T84). The project involves a wide variety of experimental techniques – including tissue culture to grow cells; end-point PCR to determine RNA expression; western Blotting to determine transporter protein abundance; immunofluorescent imaging to determine transporter cellular localization.
The overall aim of the project is to identify which intestinal cell line is the best model for future MCT1 and SMCT1 investigations. This will then eventually enable us to better understand the regulation and physiological function of these crucial colonic epithelial transporters.
Nitrogen is one of the key mineral nutrients which plants need to grow. In particular, nitrogen is required for the synthesis of amino acids, proteins, nucleic acids and plant secondary compounds and has also an important osmotic function to support cell expansion and growth. Less is known about the function of nitrogen in controlling root water uptake. There has to be some control through nitrogen, as the uptake of nitrogen and water, being the solvent for the taken-up nitrogen, has to be coordinated. Previous work in our group on barley has shown that low nitrogen supply increases the ratio of water-absorbing root surface to water-losing shoot surface considerably. This is associated with a significant decrease in the water uptake (hydraulic) properties of roots. This decrease is most likely facilitated through aquaporins, water channels in the membrane of cells. This needs to be tested.
In this project, we will grow barley plants under nitrogen-sufficient (control) and limiting (low-N) conditions and compare the expression of a set of candidate aquaporins between treatments. Expression will be quantified using qPCR. Hydraulic properties of roots will be analysed using exudation and pressure chamber measurements.
This is a combined plant physiological /molecular/environmental science project.
Dr Wieland Fricke
Plants show a significant loss of water through ‘night-time transpiration’ in the dark. This amounts to about 10% of the rate of day-time water loss. We do not know why plants loose water through the night, as they cannot do photosynthesis then, nor do we know whether most of water is lost through stomata or the waxy cuticle covering leaves. Also, we do not know whether the nutrition of plants affects nighttime transpiration.
In this project, we will make online (laptop) continuous recordings of nighttime and daytime transpiration rates of barley plants which have been grown under nutrient-sufficient conditions (‘control’) or on nutrient solution low in the key macronutrient potassium (low-K treatment). Using detached leaves, we will also determine the water transport properties of the cuticle. Together this will allow us to conclude whether most of the water lost during the night is lost through the cuticle or through stomata. In addition, we will harvest some plants during the day and night period to test whether the gene expression level (qPCR) of water channel proteins (aquaporins), which facilitate most of water transport across the membrane of root cells, changes during day and night and in response to the low-K treatment. This is a combined molecular/plant physiological/environmental sciences project.
Plants need to adjust the rate of water uptake through the root system to the rate of transpirational water loss through the shoot. It is not clear how this regulation is achieved.
The traditional view would be that such a regulation is achieved through the shoot, by means of adjusting the stomatal pore size, yet the discovery of aquaporins (AQPs), membrane water channels, has opened the possibility that this regulation can also be achieved through AQPs in roots. Barley plants which are about two weeks old have five to seven roots. When most of these roots are removed, plants still take up almost as much water as before removal of these roots. We don’t know how barley plants do this, but the most likely possibility is that the activity and possible gene expression of AQPs in the remaining roots is increased. This will enable more water uptake per unit remaining root surface.
The aim of project is to test this idea by removing all but one of the roots of 14d-old barley plants and then measure changes in plant water flow and AQP gene expression (qPCR) during the following two days. This is a combined molecular/plant physiological/environmental sciences project.
River connectivity is important for maintaining migratory fish populations, natural water and sediment movement, and diversity of instream habitat for macroinvertebrate communities and other aquatic organisms. Most rivers in Europe, including Ireland, have large numbers of physical structures such as culverts, dams, weirs, bridge aprons and fords that can hinder or prevent such movement (Fig. 1).
Fig. 1: Some examples of barriers in Irish rivers
An ongoing research project (Reconnect) in the School of Biology and Environmental Science is investigating the distribution, types and impacts of such obstacles in Irish rivers and is developing a methodology for prioritising their modification or
removal. The summer project placement will contribute to this research effort and the student will be part of a multidisciplinary team.
The student will take part in the mapping of barriers in a number of sub-catchments in Ireland. Before going into the field a desk study will be carried out using a variety of maps and images to locate potential barriers. This will be followed by field work to verify the location and types of barriers, and to carry out a series of measurements on them to draw conclusion about passability and their physical and ecological impacts. The data collected will be used to report on the types and distribution of the barriers in the study sub- catchments and where the greatest risk to connectivity lies.
Dr Rainer Melzer
We are interested in the genetic control of flower development in Arabidopsis thaliana. Of particular interest are genes that may be important for the robustness of floral organ numbers, e.g. that ensure that always four and not a more variable number of petals develop. We are currently creating mutants in several candidate genes that might be involved in this process using CRISPR-Cas9. You will be involved in generating mutant plants and/or in a detailed molecular and phenotypic analysis of CRISPR-Cas9 induced mutations. Techniques involved may comprise DNA isolation, PCR, molecular cloning, sequence analysis, plant transformation and morphological analyses.
Lab homepage: https://ucdflowerpower.org/
Dr Rainer Melzer Co-Supervisor: Saoirse Tracy
Phenotypic analyses are of eminent importance for developmental genetics and crop plant research. However, detailed and quantitative phenotypic analyses still pose a significant problem especially for small and not easily accessible structures. In this project, we aim to establish X-ray computed tomography for phenotypic analyses of Arabidopsis thaliana
flowers. We will study to which extent larger morphological but also cellular structures can be visualized using a CT scanner and how the performance compares to microscopic analyses. A number of important floral traits will be measured and comparison between wild-type and mutant plants will be conducted.
Lab homepage: https://ucdflowerpower.org/
Dr Simone Ciuti
Inter-individual variability in animal behaviour within wild populations is an important component influencing their resilience to external perturbations such as human disturbance, climate and habitat change. When such differences are consistent through time, we typically refer to them as personality traits. Personality differences are indeed a widespread phenomenon throughout the animal kingdom, with important consequences for species ecology and evolution. This project aims to test the hypothesis that behavioural traits recorded at captures (e.g. reaction to capture, vocalization, and behaviour at release) in fallow deer fawns of Phoenix Park (Dublin) are a good proxy for their personality. The student will join the Phoenix Park capture team during the second half of June 2018 and will collect behavioural data on 100 fawns during their captures. The student will then simulate a tourist walking in the park (late June – mid July 2018) and will approach fawns during their first weeks of life multiple times to classify them along a bold-shy continuum based on their (repeatable) behavioural reaction (e.g. immobile, moves, flees). The project aims to test whether behavioural metrics collected during capture are good predictors of one personality trait (i.e. consistent reaction to human presence). The ultimate goal of the study is to define a protocol for collecting personality traits in wildlife during capture.
Dr Simone Ciuti
Inter-individual variability in animal behaviour within wild populations is an important component influencing their resilience to external perturbations such as human disturbance, climate and habitat change. When such differences are consistent through time, we typically refer to them as personality traits. Personality differences are indeed a widespread phenomenon throughout the animal kingdom, with important consequences for species ecology and evolution. Personality has been shown to be moderately heritable in species’ populations, but very little research have been conducted in the wild. Provided that we suspect that personality traits between mothers and fawns are correlated, this project actually aims to test this hypotheses and estimate the strength of such a relationship (i.e., weakly vs. strongly correlated). The student will join the Phoenix Park (Dublin) capture team during the second half of June 2018 and will collect behavioural data on 100 fawns during captures (e.g. reaction to capture, vocalization, and behaviour at release). The student will then simulate a tourist walking in the park (late June – mid July 2018) and will approach multiple times fawns during their first weeks of life as well as their mothers to classify both of them along a bold-shy continuum based on their (repeatable) behavioural reaction (e.g. immobile, moves, flees). The pair mother-fawn is expected not necessarily to react the same way, with potential variation in reaction time, vigilance and escape behaviour. The project aims at testing how robust is the correlation between the personality of fawns collected at capture and during the disturbance experiments (i.e. consistent reaction to human presence) with the personality of mothers.
Dr Simone Ciuti
Obtaining correct abundance estimates to model animal population dynamics is a key goal in wildlife management and conservation. For many wildlife species, however, gathering such basic data can be extremely challenging because of elusive animal behaviour, double counting, and incorrect use of sampling techniques. Distance sampling certainly is one of the most efficient techniques developed to estimate wildlife density and abundance, but it has precise assumptions that must be met to avoid misleading results. Such assumptions are hardly met when sampling deer populations, because researchers often carry out their distance sampling surveys along tracks instead of randomly distributed transects. The goal of this research is to estimate the population size of the fallow deer in Phoenix Park, Dublin. The park is fenced, deer are habituated to human presence, terrain is accessible and visibility is good (i.e. no dense forest), which are the typical conditions of city park hosting deer populations worldwide. Visual counts performed every year estimate the population around 550 heads, and new estimates of this population will be available during winter 2017-2018. Aim of this study is to test whether visual counts usually carried out in the park are correct, under- or over-estimated when compared to distance sampling estimates. The student will apply two different distance sampling counts, i.e., i) walking and observing deer along randomly distributed transects, and ii) observing deer from randomly distributed sampling points. The student is expected to spend at least 15 days in the field: this can be adjusted, however, depending on the sampling design defined by the student and the supervisor at the beginning of the research. The study aims to train a student on a very efficient sampling technique and to establish a new protocol to count deer in city parks.
Drs Tasman Crowe and Paul Brooks
Human activities are imposing a wide range of pressures on marine ecosystems, including localised pressures such as pollution by nutrients and other contaminants and global pressures associated with climate change. These can have a wide-ranging impacts on the biota present and the functioning of the ecosystems of which they are a part. Management measures are available to limit effects of particular pressures, but these measures may themselves cause unanticipated impacts. A greater knowledge of impacts of pressures and the measures taken to limit them is required to inform effective environmental management.
In this project, you will join a research group investigating impacts of a range of human activities in Dublin Bay. Two alternative topics are possible: 1. Characterising communities of benthic invertebrates in conjunction with fine-scale sampling of water quality to test small scale variation in effects of water quality on ecosystems; 2. Testing effects of artificial structures on marine biodiversity and ecosystem functioning. Artificial structures such as sea-walls and breakwaters are increasingly being built to counter the effects of climate change, including sea-level rise and storm surges. These structures provide habitat for marine species, but often support a less diverse community than natural analogues. In this project, you would sample biota and ecosystem processes on a range of artificial structures and nearby natural shores to help identify features of structures that could be incorporated into eco-engineering solutions to reduce the negative impacts of artificial structures on marine ecosystems.
Dr Laia Comas Bru
The oxygen isotopic composition (18O) of many palaeoclimate archives (e.g. cave and lake carbonates, tree-ring cellulose), is used frequently as a proxy for past air temperature, precipitation amount and changes in atmospheric circulation patterns (e.g. the North Atlantic Oscillation). A fundamental assumption that underpins such palaeoclimate reconstructions is that the relationship between 18O and the climate parameter (e.g. air temperature) remains constant through time at the location of interest. However, a recent study (Comas-Bru et al., 2016) has shown that migrations of the Azores high sea-level pressure centre results in a non-stationary behaviour of this relationship in some European mid-latitudes. This unexpected result seriously compromises the calibration of some proxy- based palaeoclimate reconstructions. A much better understanding of how the prevailing circulation pattern influences atmospheric moisture transport is therefore crucial to the value of these palaeoclimate reconstructions as test data for ocean-atmosphere general circulation models. This project aims to provide a rationale for why the 18Op-climate relationship is affected by the persistence of atmospheric circulation patterns such as the North Atlantic Oscillation (NAO) and the East Atlantic (EA) pattern using Lagrangian back- trajectory analysis methods.
Assoc. Prof. Eoghan Holohan
Caldera volcanoes are enclosed topographic depressions formed by collapse of rocks into an underground magma chamber. Usually, such collapse occurs during a volcanic eruption and it is thought to involve the subsidence of a piston-like mass of rock along cylindrical or conical ring faults. A problem for trying to understand the factors that control how such a collapse develops is that they mostly occur underground and are thus only indirectly accessible. Consequently, many recent studies have involved numerical (computerised) or analogue (sand-box) simulations of the collapse process. These enable an insight into the inaccessible parts of the caldera system and a testing for possible controlling factors on the collapse. In this project, you will use analogue simulations to test the effect of varying the magma chamber shape on the initiation and geometry of caldera collapse. The model caldera morphologies will be quantified in 3D by using structure-from-motion photogrammetric techniques. You will systematically vary the plan-view shape of the magma chamber and the cross sectional shape of the magma chamber roof. You may also explore the influence of other parameters as you think appropriate/interesting. Finally you will compare your experimental results to data assembled on natural calderas.
Dr. Fergal O'Reilly
Much work has been done over the past decades to develop an understanding of the million degree fireballs produced when high peak power lasers interact with matter. These plasmas produce copious amounts of soft x-rays, which are becoming increasingly important for nanopatterning and materials and biological imaging. However little of this work has concentrated on plasmas with diameters of less than 20 microns, which are required to make efficient high resolution soft x-ray microscopes. In this project the student will work with and learn about high power lasers, optics, soft x-ray spectrographs and imaging systems, and vacuum systems, in cutting edge experiments aimed at tiny plasma characterisation.
Assoc. Prof. N.V. Buchete (UCD School of Physics & UCD Institute for Discovery)
Modern high-performance computational (HPC) molecular biological physics studies, including studies of protein-protein interactions, have to overcome challenges due to the separation between the relevant physical time-scales (e.g., for protein folding or binding) and the accessible time-scales of simulation. We developed recently new algorithms based on Coarse Master Equations (CME), where the underlying configuration space is discretized into a network of Markovian states, such that the mean lifetime of each state is much larger than the transition time between the states. This project (suitable for 1 or max. 2 students working separately, or as a team) will involve learning about modern Markov-based statistical methods, running simple Matlab codes for analysis, and/or designing and running simple molecular dynamics (MD) simulations using the CME-based formalism, using Linux workstations and supercomputing clusters, with the aim to improve the speedup of typical MD runs. Applications will study simple MD trajectories of diabetes or Alzheimer's diseases Abeta peptides for understanding their conformational dynamics and their propensities to form amyloid nanofibrils.
Dr James Rice
The project centres on the development and testing of plasmonic active substrates formed on polymer substrates. The aim being to study different material designs with the aim of enabling tuneable localised plasmonic resonances, achieved through mechanical displacement of the substrate. This project is experimental, centered on materials design and development.
Dr James Rice
The project aims to apply plasmonic substrates as platforms to detect small molecules through surface enhanced infrared absorption/scattering and Raman scattering methods. Thereby enabling a molecular fingerprinting of the molecules under study. The project will be experimental, centering on the application of laser based spectroscopy and FTIR instrumentation.
The following information is vetted and provided by the American Association of Collegiate Registrars and Admissions Officers (AACRAO) on the Electronic Database for Global Education (EDGE).
|Letter Grade||Percentage||Ranking||U.S. Equivalent|
|A+/A/A-||70 - 100%||First Class Honours||A|
|B+/B/B-||60 - 69%||Second Class Honours Upper||B+|
|C+/C/C-||50 - 59%||Second Class Honours Lower||B|
|D+||45 - 49%||Third Class Honours||C+|
|D/D-||40 - 44%||Pass||C|
|F||0 - 39%||Fail||F|