Permanent URI for this collection
Browse
Recent Submissions
Publication Linear regression for estimating bladder volume with voltage signals(IEEE, 2018-11-08) Dunne, Eoghan; Santorelli, Adam; McGinley, Brian; O'Halloran, Martin; Porter, Emily; European Research Council; Horizon 2020; RESPECT; FP7 People: Marie-Curie ActionsUrinary incontinence is a common condition that can severely impact the lives of those who have it. Bladder volume monitoring solutions that exploit the electrical differences of different tissues in the pelvis have the potential to help medical personnel in the decision-making process with urinary incontinence. In this work, we investigate linear regression as a means of assigning bladder volume to the measured voltage values. We found that linear regression outperforms the previously studied machine learning regression algorithms by nearly a factor of 4. This linear regression approach is also more effectively able to handle volumes outside the training boundaries in comparison to previous work in the field. More work is needed to further improve the estimate of bladder volume based on the voltage signals, especially at high noise levels.Publication Multiclass SVM for bladder volume monitoring using electrical impedance measurements(IEEE, 2018-11-08) Santorelli, Adam; Dunne, Eoghan; Porter, Emily; O'Halloran, Martin; European Research Council; Horizon 2020; RESPECT; FP7 People: Marie-Curie Actions; Irish Research CouncilUrinary incontinence is a common condition that impacts the quality of life from those who suffer from it. Electrical impedance measurements offer the potential for a non-invasive low-cost solution to monitor changes in the bladder volume. This work focuses on using a multiclass support vector machine (SVM) algorithm to classify the fullness of the bladder into three states; not full, full, and a boundary class. This paper applies this machine learning algorithm to both simulation and experimental data. The SVM model uses the recorded voltages from electrical impedance measurements as features, is trained and optimized using a Bayesian Optimization approach, and then 10-fold cross-tested to obtain a generalized error. This paper demonstrates that simulation data with a signal-to-noise ratio of 40 dB, and experimental data from a pelvis phantom, can be perfectly separated into the three classes defined above.Publication Surface chemistry and linker effects on lectin-carbohydrate recognition for glycan microarrays(2012) Kilcoyne, Michelle; Gerlach, Jared Q.; Kane, Marian; Joshi, Lokesh; |~|Glycan microarrays are an increasingly utilised tool for analysis of protein-carbohydrate interactions and a variety of glycan-containing molecules and slide chemistries have been used to array carbohydrates on microarray surfaces. Slide surface chemistry can have significant impact on the ligand presentation, background noise, spot size and morphology and reproducibility of the arrayed molecules, which in turn impacts upon lectin-carbohydrate recognition. The linker used to attach the carbohydrate to the molecular scaffold is another variable in ligand presentation. To evaluate these effects, three different microarray surface chemistries were arrayed with the same mono-and disaccharide neoglycoconjugates and natural glycoproteins and incubated with four well-characterised plant lectins. Analogues of three monosaccharide neoglycoconjugates, with two common linkers each, were included in the test group to evaluate the linker effect on lectin recognition. Based on lowest background noise, expected lectin-ligand interaction, good spot morphology and best reproducibility, the three-dimensional hydrogel slide surface proved most suitable for lectin interrogation of carbohydrate ligands, and the more flexible phenylisothiocyanate linker afforded greater recognition of the carbohydrates by the relevant lectins.Publication Upregulation of PSCDBP, TLR2, TWIST1, FLJ35382, EDNRB, and RGS12 gene expression in human myometrium at labor.(Sage, 2008) O'Brien, Margaret; Morrison, John J.; Smith, TerryThe regulatory mechanisms underlying myometrial smooth muscle contractility during labor are poorly understood. The authors therefore investigated the transcriptional profile of the changes that occur in the human myometrium at term pregnancy when compared with that at labor. Microarray technology was used to identify differentially expressed genes in human myometrium at labor. Real-time fluorescence reversetranscriptase polymerase chain reaction (RT-PCR) was subsequently performed to verify the microarray data. Semiquantitative RT-PCR, Western blotting, and microscopy methodologies were also used. Certain novel genes were found to be upregulated in human myometrium at labor. Of these, PSCDBP, TLR2, TWIST1 , FLJ35382, andRGS12 have not been previously characterized or identified in human myometrium. EDNRB is the other novel labor-associated gene whose reported expression is also upregulated at labor. All 6 genes were expressed on human myometrial smooth muscle cells. These novel upregulated genes are involved in multiple pathways that may be associated with a variety of cellular processes including inflammation, transcriptional regulation, and intracellular signaling.