DCBIOSTEC 2017 Abstracts


Full Papers
Paper Nr: 1
Title:

Regulatory T cell Development in the Human Thymus - A Comprehensive Approach Combining Genome-wide Analysis and Single-cell Protein Expression by Computational Flow Cytometry

Authors:

Yumie Tokunaga, Helena Nunes-Cabaço, Ana Serra-Caetano, Henrique Machado, Catarina Godinho-Santos and Ana E. Sousa

Abstract: The immune responses need to be tightly controlled to avoid harmful effects. T cells are key players to orchestrate this immune process. There is one T cell subset called regulatory T cells (Treg) defined by the expression of the transcription factor forkhead box P3 (FOXP3), which devoted to suppress immune responses. They can develop either in the thymus, the organ where T cells are produced, or be generated during immune responses. Thymic Tregs are considered particularly important to ensure self-tolerance and prevent autoimmunity. There are very few data regarding the factors that determine the Treg lineage commitment in the human thymus, as well as those that contribute to their maintenance after leaving the thymus as naïve Tregs. Naïve Tregs are known to continuously replenish the memory fully suppressor Treg pool, but the mechanisms involved in their maintenance throughout life are largely unknown. Our main objective is to investigate these processes by using next-generation sequencing (NGS) and computational flow-cytometry approaches. The currently available NGS data from human thymocytes are very limited. Additionally, flow-cytometry analysis has been mainly done based on a sequential gating strategy, which only focus on cell populations identified by pre-defined cellular markers. An unbiased approach will be more effective for exploring unknown developmental stages. Importantly, flow-cytometry generates multi-parameter protein expression profiles at the single-cell level. Applying computational analysis to these single-cell high dimensional data will provide relevant new relevant insights. This study is expected to significantly improve our understanding of human Treg development and homeostasis, with implications for tolerance induction and autoimmune diseases.

Paper Nr: 2
Title:

Health Informatics for Paediatric Ophthalmology - Designing Useful, Usable Information Systems

Authors:

Maria S. Cross, George W. Aylward and Jugnoo S. Rahi

Abstract: Electronic medical records (EMRs) are at the core of a recent movement towards evidence-based healthcare in many countries. In the United Kingdom, there is a target to have a paperless National Health Service (NHS) by 2020. However, a failure to understand the unique care environments of clinical specialties and to appropriately customise EMR system design threatens the delivery of any benefits. Paediatric ophthalmology is a unique field facing particular challenges in EMR adoption and data application. The heavy use of imaging and diagrammatic documentation is difficult to replicate electronically, as is the flexibility required to meet the demands incurred by the varying ages, developmental stages, and visual needs of each patient. Consideration of such requirements is essential throughout the user-centred design of effective health information technology systems. However paucity in the evidence base surrounding EMR design methodologies, applications and evaluation within paediatric ophthalmology hinders technological development and application. Therefore, within this doctoral research, we apply and evaluate a user-centred approach to health information technology development within NHS paediatric ophthalmology.

Paper Nr: 4
Title:

Biological Network Modelling and Pathway Analysis

Authors:

Ansam Al-Sabti, Mohamed Zaibi and Sabah Jassim

Abstract: ABSTRACT The search for disease-specific biomarkers for di- agnosis, illness monitoring, therapy evaluation, and, prognosis prediction is one of the major challenges in biomedical research. It has long been that diseases are rarely caused by abnormality in a single protein, gene or cell. But by disorder of different processes man- ifested by intracellular network of interactions be- tween the molecular components in such biological systems. Despite the popularity of biological network anal- ysis methods and increasing use for identifying genes or pathways (groups of genes) that contribute to diseases and other biological processes, impor- tant topological and network information are hardly used in ranking/assessing the relevance of the path- ways. Often, gene expression values and confidence score/strength of interactions are not considered when scoring/ranking the resulting pathways. The research presented in this paper focuses on two different, but closely related areas in Bioinfor- matics: developing new approaches for biological network analysis, and improving the identification of disease biomarkers. The inclusion of topological weight and expression level in the calculation of path- ways score is expected to facilitate the identification of the pathways that most relevant to pathophysiolog- ical processes.

Paper Nr: 5
Title:

Wearable Technology in the Study of Raynaud’s Phenomena - Ascertainment of the Potential Impact of Wearable Technology on Raynaud’s Phenomena Utilizing Data

Authors:

Isobel Taylor, Heitor Alvelos, Susana Barreto and Pedro L. Granja

Abstract: Can wearable technology aid in the study of Raynaud’s Phenomena? Raynaud’s phenomena (RP) affects 3-20% of the Population worldwide. RP is the vasoconstriction of the microvascular system in the extremities, such as the digits, in response to cold exposure or emotional stress. The prime quotidian problems for sufferers are: ischemia to the extremities, the pain this causes, loss of fine pincer movements and the aesthetics. The initial symptom of an RP attack is often numbness; this is not always noticed by the individual with RP.The faster the attack is identified the sooner the individual can attempt to re-establish blood flow, therefore shortening the time the tissue has been starved of nutrients. Current options include Pharmaceutics which have side effects from headaches to bleeding in the brain along with problems in cross drug interactions. Whilst available wearables such as ski gloves and heat focused garments limit function and dexterity they also have a lack of data of the impact these have. The presupposition is that a wearable data collection device could create a database to aid in the understanding of the condition itself as well as any wearable or pharmaceuticals impact. The PhD research continues from the researcher’s MRes in Digital Media that questioned whether an RP attack could be detected with a temperature sensor. The study showed positive results which included taking the skin surface and environmental temperatures of predicting attacks initiating from data changes. The research will be advanced by focusing on data gathering, for an accurate device, improving the calculations involved. The aim is to create a device/ system to extract data from the skin concomitant with external data including weather reports and geo location incorporated through exploitation of a smart phone for data gathering and assessment over time. Interest also lies in how a designer can impact the implementation of the research. The research will centre on Primary Raynaud’s Phenomena, PRP, in the fingers with the potential this could be used in cases of Secondary Raynaud’s Phenomena, SRP, and other effected extremities.

Paper Nr: 6
Title:

Towards Novel Methods for Effective Transfer Learning and Unsupervised Deep Learning for Medical Image Analysis

Authors:

Mijung Kim, Jasper Zuallaert and Wesley De Neve

Abstract: Increasing attention to the deep learning applications comes to the medical image analysis as well. Recently, Google published the paper on detecting the diabetic retinopathy using their deep learning approach. During our doctoral research, we will mainly focus on dealing with medical image dataset from the deep learning's perspective. In particular, due to its sensitive nature ans sparsity, medical image dataset does not show high performance as other images do such as ImageNet. By making use of transfer learning and unsupervised learning techniques we will investigate in increasing the effectiveness our models in medical image dataset. As for the first step, we challenge to the breast cancer mammography dataset using Inception V4 to diagnose the lesions. After applying data augmentation methods we will move on to unsupervised learning approach to overcome small size and unlabeled dataset. Ultimately, we will develop a unsupervised deep learning network with transferable discriminator maintaining high effectiveness.