Abstracts Track 2022

Nr: 5
Title:

Virtual Walking
Authors:

Roger Abächerli
Abstract: A major common complication that characterizes paraplegic patients besides the inability to walk is chronic, neuropathic pain. This type of pain is not easily amenable to treatment. Various methods have met success in a limited number of patients, but the majority of patients still have to contend with this type of chronic pain, often throughout their life. The “mirror”-therapy has proven to be efficient in amputees suffering from neuropathic chronic pain similar to that of paraplegics. Consequently, a system called “VitualWalk” has been developed that enables the paraplegic to mimic healthy walking in a controlled environment by analogy with the mirror therapy. “Virtual walk” combines a video layer taken offline simulating a natural environment such as a walk in a forest, a video of selected moving legs acquired from healthy subjects, a video from the head and thorax of the patient sitting in a chair taken in real-time, ambient audio recordings and a seating surface integrated into the electronic wheelchair carrying the patient’s pelvis moves in a way closely mimicking the pelvic motions characteristic of healthy walking. The main challenge of “Virtual Walk” was to keep all free visual layers and the haptics in good synchrony as latency is known to create headacke, which is exactyl the opposite of what the system should create. The synchronization between the legs recording, the cycle phase and the seat’s motions is essential to the robustness of the illusion and hence to the postulated therapeutic value of the treatment. Furthermore, we created a digital-twin of the pelvis' motion of healthy people in order to create a ideal mechanical motion of the seating surface. This psycho-physical illusion of a walking state has a therapeutic effect whereby the neuropathic pain was reduced in several patients in a pilot feasibility study. Therefore, the “Virtual Walk” system offers another therapeutic option, as a last option, predicated on the relationship between pain and perception that is less explored than the conventional methods.
Nr: 7
Title:

Detecting the First Stages or Resurgence of Lymphedema using a Connected Device: Preparing a Clinical Study by Designing a Digital Twin
Authors:

Nicolas Bousquet
Abstract: Testing and quantifying (or even improving) the capabilities of a connected device, through a pre-clinical or clinical study, can be particularly time consuming and costly, especially in the healthcare field. The development of simulation models of the device including a part of the human body, the device itself and its electronics, as well as the system allowing the monitoring of measured and wirelessly transmitted quantities can drastically decrease this cost by improving patient selection and providing clear paths for design improvement. We propose to explain a methodology designed for this purpose, which uses statistical approaches for uncertainty handling coupled with geometrically translated biomechanical considerations, through the study of a connected device aiming at measuring the arm diameter of a patient likely to develop post-masectomy lymphedema following breast cancer (more precisely: axillary lymph node dissection). This pathology can be extremely disabling and quasi-irreversible. It is wanted to alert a patient of any manifestation or resurgence in order to prevent its aggravation. The management of the global measurement noise mainly due to physical movements and the positioning of electronic circuits is key to achieve a use that can allow the detection of a weak signal and a quick medical response, without the ergonomics becoming problematic for the patient. For this reason, a simulator of the entire physical process (called "digital twin") has been developed for this purpose, in a hardware-in-the-loop framework. Statistical calibration of its input parameters, by stochastic inversion and using sensitivity studies, led to establish one or more measurement protocols allowing to capture the signal on a mobile device (phone or tablet) and to detect signal breaks that are physically significant. The measured signal makes it possible to report quickly on the worsening of the patient's condition and to warn the therapists within a very reasonable period of time. The general methodology of this work seems to us to be easily reproducible in the preparation of clinical studies of other types for connected devices, which aim to develop measurement protocols limiting the often significant cost of such studies. For this talk we rely on the following article: L. Béthencourt et al., "Guiding Measurement Protocols of Connected Medical Devices Using Digital Twins: A Statistical Methodology Applied to Detecting and Monitoring Lymphedema," in IEEE Access, vol. 9, pp. 39444-39465, 2021, doi: 10.1109/ACCESS.2021.3063786. This article is provided as a complementary material for this presentation, in accordance with its publishing licence.
Nr: 9
Title:

UP SIBOL: The Role of Collaborative Innovation in the Philippine COVID-19 Response
Authors:

Miguel Sandino O. Aljibe, Edward H. Wang, Philip B. Fullante, Miguel J. Paraiso, Charleston Dale M. Ambatali, Maria Teresita B. Aspi, Catherine S. Co, Louis M. Danao, Leslie Joy L. Diaz, Emmanuel P. Estrella, Samuel Arsenio M. Grozman, Maria Antonia E. Habana, Geohari L. Hamoy, John Richard E. Hizon, Manuel C. Jorge II, Eduardo R. Magdaluyo, Jose Donato A. Magno, Portia F. Marcelo, Michelle Cristine B. Miranda, Prospero C. Naval, Jr., Nathaniel S. Orillaza, Jr., Jason Pechardo, Marc D. Rosales, Luis G. Sison, Magdaleno R. Vasquez and Jr.
Abstract: The Philippine healthcare system has historically been plagued by the unavailability and incompatibility of medical devices for the local population, especially those used in surgery. In 2015, a multisectoral effort through the Philippine Biomedical Device Innovation Consortium (PBDIC) was organized to encourage the local development of medical devices. Continuing this effort in 2019, the University of the Philippines Manila (UP) and the Department of Science and Technology (DOST) started the process of establishing a Surgical Innovation and Biotechnology Laboratory, which was eventually named UP SIBOL. However the onset of the COVID-19 pandemic forced the nascent organization to adapt in order to address the rapidly evolving health crisis. The objective of this paper was to describe the novel experiences of UP SIBOL innovators during the pandemic. Focus was given on the use of cloud technologies that enabled design and fabrication activities in spite of quarantine restrictions. Qualitative data on these experiences were extracted from interviews and review of the organization’s internal documents. Ten device research projects were initiated by UP SIBOL for COVID-19 response. These projects were divided into four categories: disinfection, personal protective equipment (PPE), telemonitoring, and support devices. Several devices have been deployed in the COVID-19 wards of the Philippine General Hospital like Sanipod, a disinfection cubicle for hospital staff, and myBESHIE, a telepresence robot. Another project, a breath responsive powered air purifying respirator (PAPR) is under evaluation by local manufacturers for mass production. Several other projects are in the final stages of prototyping and are slated to be deployed to assist health workers. The completion of the above mentioned projects suggests that the UP SIBOL experience can serve as a framework for the use of collaborative innovation in response to novel crises.
Nr: 13
Title:

Calling Arbitrarily Nested Structural Variants using an Unambiguous Representation Scheme
Authors:

Markus R. Schmidt and Arne Kutzner
Abstract: Structural variant (SV) calling identifies differences between a ‘reference’ and a ‘query’ genome, using the query genome’s sequencing reads and the reference genome’s assembly. Callers have evolved from detecting simple alterations, like mutations, to increasingly complex ones, like inversions and translocations; however, the step to detecting nested variants has not been completed yet. Our work focuses on nested SVs that state-of-the-art callers fail to resolve. We reveal that the callers’ issues arise from twofold conceptual ambiguities in their SV representation: i) The same alteration can be expressed in different ways, and ii) different alterations can be represented in the same way. We resolve these ambiguities using a graph-based model. For this model, we translate break-end pairs, which are detected from individual reads, into graph edges. Then a line-sweep over the graph’s adjacency matrix is used to cluster nearby edges. Each cluster of edges finally represents the consensus of all reads for one SV. For two yeast genomes, we show that our graph model can wholly represent a query genome as SVs to a reference genome. Using simulated Illumina and PacBio reads, we demonstrate that we can generate instances of our graph model with high accuracy and recall. With our approach, we extend the state-of-the-art SV representation scheme to a model that can unambiguously represent all kinds of variants, which is a prerequisite for obtaining a comprehensive overview of SVs.
Nr: 14
Title:

Novel Interferon-sensitive Genes Unveiled by Correlation-driven Gene Selection and Systems Biology
Authors:

Riccardo L. Rossi, Cristina Cheroni, Lara Manganaro, Lorena Donnici, Valeria Bevilacqua and Raffaele De Francesco
Abstract: Interferons are key cytokines involved in alerting the immune system to viral infection. After interferons stimulation cellular transcriptional profile critically changes, leading to the expression of several interferon stimulated genes (ISGs) that exert a wide variety of antiviral activities. Despite many ISGs have been already identified, a comprehensive network of coding and non-coding genes with a central role in IFN-response still needs to be elucidated. We performed a global RNA-Seq transcriptome profile upon IFN treatment of two HCV permissive human hepatoma cell lines and defined a network of genes whose coordinated modulation plays a central role in IFN-response: we then processed the results with a combination of correlation network analysis and "inverse" gene-ontology functional genomics approach to get rid of all already annotated genes. This allowed us to sctratch under the surface to unveal previously uncharacterized interferon stimulated genes. Our study adds molecular actors, coding and non-coding genes, to the complex molecular network underlying IFN-response and shows how systems biology approaches, such as correlation networks, network’s topology and gene ontology analyses can be leveraged to this aim.
Nr: 15
Title:

Database Development of MedScrab: An Interactive Game to Improve Medication Information Recall
Authors:

Don Roosan
Abstract: Background: Healthcare system is moving toward a patient-centric model. As a result, patient education plays a vital role in all aspects of care. Poor health literacy has a direct association with poor health outcomes, including hospitalization, mortality, and chronic disease management(Pignone & DeWalt, 2006)(Aboumatar et al., 2013). Despite the detailed prescription patient handout provided, critical gap between patient’s understanding of medication information recall during adverse events still exists. Most patients do not read the lengthy patient leaflet, which resulted in a lack of understanding of medication information. In multiple studies, visual aids have shown to impact information recall significantly compared to text alone. (Farrell et al., 2014) (Katz et al., 2006). Therefore, to address the knowledge gap, we proposed an interactive game-based application that can help to improve patient’s recall of medication information. We hypothesise that by recalling crucial medication information, patients may be prone to higher adherence, lower adverse event rates and overall better health outcomes. In the game application, we first develop a database framework to organize and integrate crucial medication information. We focused on cardiovascular and mental health medication for the prototype. Method: We designed a conceptual data model for a mobile game application. The data model includes six entities: drug class, drug information (e.g., generic name, brand name, keywords), drug information type (e.g., indication, warning, adverse effects, interactions, and patient counselling), DrugQuiz (e.g., quiz question, hint) DrugQuizOption, and QuizType (i.e., sorted based on the level of the game level). We then created a front-end web form and populated 2100 medical information related to 84 cardio and mental health drugs. The crucial drug information was reviewed by clinical pharmacists. Result: We successfully created a database for Medscrab, an interactive mobile game application that allows the participants to pick their drug of choice to play the game. The database contains a total of 2100 information derived from 84 medications. Each medication has 25 pieces of information that belong to 5 different levels and 9 different quiz questions with multiple choices answers. After completion of each level, there is a mini-quiz and a final quiz at the end of each game to examine the participants’ knowledge and ability to recall displayed information. We also incorporated feedback from all team members and reiterated the design of the database several times. Conclusion: In this study, by using a relational database and ironic framework, MedScrab, an interactive mobile game application, was successfully created to educate the patient on medication information. The application is currently available on both iOS and Android systems (www.medscrab.com). In the future, we will be conducting usability and experimental design studies.
Nr: 22
Title:

Elasto-magnetic Pumps for Point-of-Care Diagnostics
Authors:

Jacob Binsley, Stefano Pagliara and Feodor Ogrin
Abstract: Microfluidic lab-on-a-chip (LOC) devices are pivotal for the progression of point-of-care (POC) diagnostics. Their effectiveness has been shown recently with the widespread advent of the lateral flow immunoassays used for rapid testing of COVID-19[1]. However, depending on passive flow generation such as capillary forces does not easily allow for prolonged operation times or constant flow rates. We have seen in recent decades, a surge in the development of integrated pumping solutions; miniature on-board pumps which can enhance the capability of POC devices outside of laboratory environments[2]. Pumping and swimming on this scale is non-trivial. Heavily damped Aristotlean mechanics must be considered and fluid flow must be generated through a non-reciprocal sequence of pump motions. We have developed a possible novel pumping solution and have fabricated and tested an experimental, elasto-magnetic micropump inspired by Purcell’s 3-link swimmer[3]. This design, consisting of 3 elasticated links made from PDMS, is coupled to an external magnetic driving field via the inclusion of a small Neodymium Iron Boron magnet. When provided with a weak, uniaxial oscillating driving field with flux density of 10s of gauss and frequency in the range of 10s of Hz, the device produces tuneable and reversible fluid flow. During our most recent investigations, we modelled this system numerically using COMSOL Multiphysics to not only verify our experiments, but to explore the device more deeply and gain an understanding of how this system may be optimised. These new numerical results will be the main subject of this talk, and offer valuable and very visual insight into bio-inspired pumping mechanisms for lab-on-a-chip POC devices. [1] Andryukov, Boris G. "Six decades of lateral flow immunoassay: from determining metabolic markers to diagnosing COVID-19." AIMS microbiology 6.3 (2020): 280. [2] Boyd-Moss, Mitchell, et al. "Self-contained microfluidic systems: a review." Lab on a Chip 16.17 (2016): 3177-3192. [3] Binsley, Jacob L., et al. "Microfluidic devices powered by integrated elasto-magnetic pumps." Lab on a Chip 20.22 (2020): 4285-4295.
Nr: 23
Title:

Continuous Monitoring of Monoclonal Antibody Breakthrough from Capture Columns using Immobilized Fluorescent Reporters
Authors:

Atul Goyal, Binh Vu, Vijay Maranholkar, Ujwal Patil, Katerina Kourentzi and Richard Willson
Abstract: In 2020, the global biopharmaceutical market was valued at over $325 billion, with monoclonal antibody (mAb) therapeutics representing the majority [1]. Currently, 68 mAb therapeutics have been approved by the FDA, with ~570 more in clinical testing for applications ranging from cancer to eczema to COVID-19 [2]. For instance, in November 2020, the US FDA issued an emergency use authorization (EUA) for casirivimab and imdevimab to be administered together to treat mild to moderate COVID-19 in adults and pediatric patients who are at high risk for progressing to severe COVID-19 [3]. Measurement of antibody concentrations is ubiquitous in biopharmaceutical process development and manufacturing. Purification of therapeutic mAbs usually involves a protein A affinity capture step, which has become a gold standard for the industry. Advancements in cell culture technology have enabled very high antibody titers which create incentives for high column loadings, but the imperfect mass transfer in capture columns can risk breakthrough of the valuable product. Conservative under-loading, however, wastes expensive protein A chromatography resin, which can cost $12,000liter [4]. Because column breakthrough of antibody in complex, UV-absorbing culture fluid cannot be readily detected in real-time, processes are conservatively designed and column capacity often is underutilized, wasting adsorbent and reducing productivity. We previously developed a fluorescence-based monitoring technology which allows mix-and-read mAb detection in cell culture fluid, which may be useful in at-line assays and in clone and culture development. Here we report the use of reporters immobilized on CNBr-activated Sepharose 4B resin for continuous detection of IgG in column breakthrough. The column effluent is continuously contacted with immobilized fluorescein-labeled Fc-binding ligands to produce an immediately-detectable shift in fluorescence intensity. The technology allows rapid and reliable monitoring of IgG in a flowing stream, without prior sample preparation. We observed significant shifts in fluorescence intensity at 0.5 g/L human IgG, sufficient to detect a 5% breakthrough of a 10 g/L load. The fluorescence intensity response at different load concentrations was used to calibrate fluorescence intensity with IgG concentration. Bibliography [1] Expert Market Research, “Global Biopharmaceutical Market,” 2020. [Online]. Available: https://www.expertmarketresearch.com/reports/biopharmaceutical-market. [2] M. K. Maruthamuthu, S. R. Rudge, A. M. Ardekani, M. R. Ladisch, and M. S. Verma, “Process Analytical Technologies and Data Analytics for the Manufacture of Monoclonal Antibodies,” Trends Biotechnol., vol. 38, no. 10, pp. 1169–1186, 2020. [3] U.S. Food and Drug Administration, “Coronavirus (COVID-19) Update: FDA Authorizes Monoclonal Antibodies for Treatment of COVID-19,” 2020. [Online]. Available: https://www.fda.gov/news-events/press-announcements/coronavirus-covid-19-update-fda-authorizes-monoclonal-antibodies-treatment-covid-19. [4] D. Stanton, “Protein A’s mAb capture efficacy offsets cost, says GE launching new offering,” BioPharma-Reporter.com, 2017. [Online]. Available: https://www.biopharma-reporter.com/Article/2017/09/26/Protein-A-s-efficacy-offsets-cost-says-GE-as-it-launches-new-offering.
Nr: 24
Title:

Peptide Minimal Model for Studying Aggregation and Fibril Formation
Authors:

Beata Szala-Mendyk
Abstract: In the living cells, the protein-protein interactions are responsible for many important processes, such as formation of protein fibrils or liquid-liquid phase separation. The protein droplets or aggregates may play important roles in living cells but they are also connected with the numerous disorders such as Alzheimer's, Parkinson's or Huntington's diseases [1]. Despite many studies, the molecular mechanism of protein and peptide aggregation is not completely understood. Also, the origin of the aggregate polymorphism has not been elucidated yet [2]. To gain insight into the molecular details of the peptide self-assembly we used molecular dynamics simulations. Due to the number of molecules involved in the peptide aggregation, it is difficult to simulate this process with atomistic resolution. For that reason, we propose simple, implicit solvent coarse-grained model which consists of Lennard-Jones super-atoms and has initially only one super-atom per residue [4,5]. This model allows for the reproducing of the different types of aggregation kinetics and also the variable aggregates structures as a function of two model parameters: inter-molecular interaction strength and chain stiffness. The simulations with this uncomplicated model reveal also the chiral fluctuations of the oligomers formed in the early aggregation stage. The addition of explicit side chain super-atoms extends the assortment of observed structures, including the fibril-like species. Our recent study indicates, that the aggregate morphology depends on the size of side chains as well as the length of the side chain-backbone bond. In our work, we present how the subsequent model elements may affect the aggregation. This work is supported by grant no.\ POWR.03.02.00-00-l026/16 co-financed by the EU through the European Social Fund [1] T. P. J. Knowles, M. Vendruscolo, and C. M. T. Dobson, Nat. Rev. Mol. Cell Bio., 2014, 15: 384-396 [2] R. Gallardo, N. A. Ranson, and S. E. Radford, Curr. Opin. Struct. Biol., 2020, 60:7-16 [3] S. Kmiecik, D. Gront, M. Kolinski, L. Wieteska, A. E. Dawid, and A. Kolinski, Chem. Rev., 2016, 116:7898- 7936 [4] B. Szała, A. Molski, Soft Matter, 2020, 16:5071-5080 [5] B. Szała-Mendyk , A. Molski, J. Phys. Chem. B, 2021, 125:7587-7597.
Nr: 32
Title:

Single-cell Mapping of microRNA Expression during Cardiac Development
Authors:

Stefanos Leptidis, Eleni Papakonstantinou, Katerina Pierouli, Thanasis Mitsis, Sarantis Chlamydas, Aspasia Efthimiadou, George P. Chrousos, Elias Eliopoulos, Emil Hansson and Dimitrios Vlachakis
Abstract: The heart is an exceptionally complex tissue, comprised from a variety of different cell types. Understanding physiological cardiac development and its relationship to the development of pathological cardiac disease, require the careful investigation of their related developmental pathways. A highly significant regulatory layer during cellular differentiation is the post-transcriptional regulation via non-coding RNAs and more specifically microRNAs. Previous microRNA transcriptomic studies in the heart were lacking in the identification of their differential expression per cell-type. Since microRNAs can target a large range of mRNAs, identifying their cell-type specific expression is necessary, both to elucidate the intricate cellular interactions and regulatory pathways, as well as the development of targeted therapeutic approaches. In this study, we are using data from single-cell small RNA sequencing (small-seq) from early embryonic cardiac progenitor murine cells. Small-seq is an adjusted single-cell RNA sequencing method, based on the smart-seq2 protocol, which allows for the selective identification of the transcriptional profile of non-coding RNAs such as microRNAs, snoRNAs, tRNAs and lncRNAs. Our approach aims to identify the expression of these populations during early cardiac development. To that end we are using isolated cardiac progenitor cells from mouse E8.5 and E9.5 embryonic hearts. These populations are selected on the basis of their expression of early cardiac transcription factors, such as Nkx2.5 and Isl1, which delineate the emergence of the First and Second Heart Field regions, during cardiac development. Unlike single-cell RNA sequencing (scRNAseq), there are no established cell-type markers nor data analysis methods in the case of small-seq. Consequently we have been developing a methodology for the identification of cell-types using their microRNA profile, coupled to their predicted targets stemming from various miRNA target prediction algorithms. Since the small-seq protocol has also yielded a number of mRNA transcripts, we are overlaying this additional modality to further increase our cell-type identification process. These data are cross-referenced with scRNAseq data in the same tissue, isolated during the same developmental stages, where the cell-types have already been identified. Through our method we have identified a small number of differentially expressed miRNAs, clustering within two main populations. Additional sublustering reveals the potential existence of sub-populations within those clusters. Furthermore we have also identified a large number of lncRNAs, as well as a small number of snoRNAs, snRNAs and scaRNAs. While subsequent analysis shows that they do not form highly distinctive clusters, there are indications of distinct expression for a number of them. Additionally, using only the mRNA transcript information acquired we were also able to identify at least 5 different clusters, stemming just from the small-seq data. Deciphering the transcriptomic landscape of microRNAs during cardiac development, along with identifying cell-types based on the relationship between their RNA and microRNA fingerprint, enables the in-depth study of the intricate regulatory interactions between cells, cell-types and different embryonic days.
Nr: 35
Title:

Detecting Predictive Signatures of Asthma in Pediatric Breath Samples using Machine Learning
Authors:

Saurav Bose, Amalia Z. Berna, Ariel Hernandez-Leyva, Aaron J. Masino, Andrew Kau and Audrey R. Odom John
Abstract: Introduction: Asthma is a common allergic disorder characterized by airway inflammation and obstruction that affects at least 300 million people and accounts for up to 1 in 250 deaths worldwide. However, currently available diagnostic tests of asthma are either invasive or require patient compliance, making them difficult to use in children and necessitating the invention of new methodologies. A small number of studies have shown preliminary success using computational tools to mine breathomics data for pediatric asthma markers as a noninvasive alternative. However, most of these studies focused on distinguishing between asthma subtypes. In this study, we collected breath samples from children with and without asthma and developed a novel end-to-end computational pipeline to (1) distinguish between the two populations and (2) identify biomarkers that correlate with asthma. Methods: This retrospective, IRB approved study included 43 children (asthmatics, n = 14) between ages 6 and 12 years whose breath samples were collected and analyzed between January 2019 and January 2020. The breath samples were processed using a GC/MS-qTOF system to identify constituting ions and create a high-dimensional dataset with 2734 features representing ion intensities. Embedded in a leave-one-out cross-validation methodology, Pearson correlation-based dimensionality reduction (CDR) followed by one of five filter-based feature selection methods – (1) CDR+ ReliefF, (2) CDR + MultiSURF, (3) CDR + ANOVA, (4) CDR + ANOVA + ReliefF, (5) CDR + ANOVA + MultiSURF was performed to create feature subsets of varying dimensionalities. These feature subsets were evaluated for their ability to discriminate between individuals with and without asthma using the XGBoost algorithm (hyperparameters tuned using Bayesian optimization). For the most optimal feature selection routine based on area under the ROC curve (AUCROC), features that were selected in at least half of the validation splits were considered discriminative. Finally, corresponding chromatogram peaks of the 17 selected features were examined to perform compound identification. Results: A feature selection pipeline consisting of CDR, ANOVA feature selection and Relief-F in combination with XGBoost performed best in reducing the data to a set of highly discriminative features. This procedure resulted in an AUCROC of 0.91, AUCPR of 0.88, NPV of 0.87, Precision of 0.83, Recall of 0.71, Specificity of 0.93 and F1 of 0.77. We also observed that the non-linear XGBoost model consistently outperformed the linear Logistic Regression model for all feature selection methods. Finally, a total of 5 compounds – pentanoic acid, cyclohexylmethane, 3-carene, toluene and L-alanine, 3-sulfo- were identified from the 17 discriminatory features selected by the algorithm. Conclusion: We developed an end-to-end system to process high-dimensional GC-MS derived data to identify the most discriminative compounds (biomarkers) present in the breath samples of children with and without asthma. Quantitative and qualitative results presented in this study compare favorably to prior work. Although the biomarkers we identified are preliminary, they have the potential for use in devices such as electronic noses which can replace the current invasive diagnostic tests for pediatric asthma. Future research is warranted to understand the origin of the discriminatory biomarkers identified here and their relation to the pathology.
Nr: 10
Title:

Data Harmonisation of Australian and New Zealand Ambulance Service Datasets
Authors:

Kathryn Eastwood, Alison Johnson, Angela Jones, Peter Cameron and Helena Teede
Abstract: Background Eight state-based ambulance services (emergency medical services; EMS) across Australia and two across New Zealand (NZ) provide emergency medical response to telephone calls for assistance for the 26 million Australians and five million New Zealanders. Because these ambulance services operate independently from one another, significant variation exists in their patient data collection methods, the variables collected and the variable definitions. This has compromised performance benchmarking, clinical audit and cross-border research opportunities and translation of research to improve patient care. Ambulance data harmonisation has occurred in the United States and United Kingdom however, to-date no data harmonisation has occurred in Australia. Objectives This study aims to compare ambulance service variables in Australia and NZ to identify opportunities and barriers for data harmonisation. Method Dataset variable lists for four ambulance services were available online, and the remainder were requested from the ambulance services. Three services provided their lists, and three did not respond to the request. Two of these used the same electronic patient care record (ePCR) system as one whose list had been provided and one variable list was not sourced. Variables were mapped to each other and several international standardized terminology systems to identify variations and similarities in variable names and definitions, and harmonisation opportunities. Results The ambulance services are at varying stages of maturity with respect to data collection techniques. Four Australian ambulance services used the same ePCR system, three used other ePCR systems, one used paper-based records and both NZ services used a single ePCR system. Only the NZ services had mapped their variables to two international standardised terminology systems (Systematized Nomenclature of Medicine – Clinical Terms (SNOMED-CT) and Logical Observation Identifiers, Names and Codes (LOINC) terms). Three main barriers to harmonisation were identified. These included the variables collected, the variable definitions and the variable naming convention. The core variables available for mapping varied and numbered from 27-69. Across the datasets, variables with similar names often had different definitions and variables that should have had different definitions, had the same. For example ‘gender’ and ‘sex’ often had the same definition, despite accepted definitions indicating that ‘sex’ refers to the biological chromosomal and anatomical distinction whereas ‘gender’ is defined as the gender to which a person identifies which can include ‘unspecified’, 'transgender/sexual', 'gender diverse’, 'pan-gendered‘, and 'inter-gender‘. Finally, the naming convention for similar/same variables differing between services. For example the suburb to which an ambulance is called could be named ‘scene suburb’, ‘city’, ‘suburb’, ‘scene location’ or even ‘From3’. Conclusions Variation exists in the variables collected by ambulance services across Australia and NZ presenting a range of barriers to data harmonization. However, the mapping in this study demonstrates that data harmonisation in Australia and NZ is possible and presents significant opportunities for improvement in patient outcomes and performance audit. It would also facilitate quality, large-scale, high-impact collaborative national and international research.
Nr: 20
Title:

Modelling of 3D Placental Cell Features using Deep Learning
Authors:

Ben Mills, James Grant-Jacob, Benita S. Mackay, Rohan Lewis and Bram Sengers
Abstract: Serial block-face scanning electron microscopy (SBFSEM) is a well-established technique for producing a sequential series of high-resolution images of a material, from which three-dimensional structures can be determined. This is achieved by incorporating a diamond knife microtome within a scanning electron microscope (SEM). The result is a "stack" of sequential SEM images that correspond to the 3D volume of the sample. Typically, a SBFSEM stack might include ~500-2000 SEM images, each with ~10nm spatial resolution in x and y, and separated in height by ~50nm in z. In our work imaging human placental tissue using SBFSEM, we currently employ human labelling of features within the SEM image stacks, which allows 3D reconstruction and analysis. Whilst human labelling can be very accurate, the amount of time needed to label a single feature in a SBFSEM stack can be several months, and hence there is a practical limit to how many image stacks can be analysed. There is therefore a clear need for alternative methods for analysis that do not require human labelling. Here, for the first time, we use a neural network to generate an SEM image when we provide the overall shape of the structures within the placental tissue. This is based on recent work in the field of generative neural networks that can be used to generate “fake” images of human faces [1] (also see the interactive website [2]). The analogy is that we are using a neural network to generate “fake” SEM images of placenta cells. This process can be repeated for any 3D cell shape, and the results so far are shown to be statistically consistent with experimental data. This early result therefore offers the tantalising prospect of using a generative neural network for data-driven modelling of 3D cell biology. The work presented here takes several novel steps. Firstly, we use a custom-written algorithm to label the boundary of the placenta cell in every SEM image in a stack. For each image, this boundary is used to make a “mask” that is labelled green outside the cell and red inside the cell. Secondly, we train a neural network to transform a green/red mask into a “fake” SEM image, meaning we can therefore generate an SEM image for any placental cell shape. Thirdly, we extend this capability into 3D, and use the neural network to generate a “fake” stack of SEM images for any chosen 3D placental cell shape. As a demonstration of a potential application of this approach, we label the position of generated blood cells in each of the generated SEM images in the “fake” stack, and our next aim is to visualise a predicted 3D capillary network for a placental villi. This work builds upon our previous work [3] (presented at BIOSTEC 2020), where we demonstrated that a neural network can be used to automatically label features of placenta cell types, such as the endothelial cells and pericytes. This was achieved through training a neural network on SEM images of placenta cells that had the features of interest already labelled by a human. [1] Karras, T. et al, “A style-based generator architecture for generative adversarial networks”, IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019, 4401-4410. [2] https://thispersondoesnotexist.com/ [3] Mackay B.S. et al, “Deep Learning for the Automated Feature Labelling of 3-Dimensional Imaged Placenta”, BIOSTEC 2020, doi:10.1007/978-3-030-72379-8_6
Nr: 21
Title:

Problems and Prospects of Creating an Automated Personalized System for Glucose Level Monitoring and Controlling.
Authors:

Kirill V. Pozhar, Nikita M. Zhilo, Evgeniia L. Litinskaia, Mikhail O. Mikhailov and Nikolay A. Bazaev
Abstract: With the development of telemedicine and personalized medicine, one of the priorities in the field of devices for the treatment of type 1 diabetes mellitus has become the development of automated systems for monitoring and controlling blood glucose level. Such systems include a personalized wireless glucose sensor, an insulin pump, and an automatic insulin infusion rate control algorithm. Due to the significant delay between insulin administration and its action, it should be administered in advance using model predictive control. One of the most topical problems of this method is the insufficient efficiency of the existing mathematical models of glucose dynamics in the body. The most effective model on which the T1DMS simulator is based, in particular, does not describe the important processes associated with the dynamics of glycogen and glucagon, as well as the effect of counterinsular hormones, including catecholamines, on the processes of glycolysis and glycogenesis. As a result, stress, fasting or physical activity situations are described incorrectly. An increase in the accuracy of describing these processes, at the same time, can lead to a complication of the model, and the more complex a model is proposed, the more difficult is its identification procedure. Another important aspect of building a glucose control system is the choice of control strategy. We have proposed a combined feedback-control strategy to control basal insulin infusion in combination with a model predictive control to respond to situations that could cause a sudden rise or fall in glucose levels. To reduce the risks caused by inaccuracies in the calculation of the recommended profile of insulin bolus delivery, it is proposed to implement its infusion through patient decision support. The collection of statistics on the dynamics of glucose, the situations that have arisen and the decisions made will allow the attending physician to adjust the therapy. One of the important aspects of the effectiveness of insulin therapy systems is their ergonomics. Modern glucose sensors are partially implantable and require regular painful replacement. The development of non-invasive glucose monitors is promising. Experimental studies of the developed optical glucose sensor based on NIR diffuse reflectance spectroscopy show that its accuracy for single glycemia measurements is insufficient. At the same time, the meter shows the dynamics of glycemia quite well, which can be effectively used to control excessive or insufficient regulation. Such a sensor, like a partially implantable one, requires periodic calibration, but its use is much less painful and more ergonomic, which makes the procedure of continuous insulin therapy more effective. Automated systems based on a non-invasive glucose sensor can also be used in wearable peritoneal dialysis machines to maintain the concentration of glucose as an osmotic agent. If in the first case the glucose concentration is maintained by the administration of insulin, in the second - by the administration of glucose. Experimental studies have shown that the accuracy of a non-invasive optical transmission glucose meter is sufficient to ensure effective maintenance of the osmotic agent. Such a sensor, in turn, can be used to indirectly control the ultrafiltration rate in automatic wearable machines. The reported study was funded by RFBR, project №20-37-90049\20.
Nr: 25
Title:

CisTarMir: A Bioinformatics Pipeline of Identifying microRNAs Enriched with cis-miR-eQTLs and Neighboring Target Genes
Authors:

Eric Huang and Yongsheng Bai
Abstract: MircoRNAs (or miRNAs) are short or small RNAs that regulate gene expression through targeting the 3' untranslated region (3'UTR) of target genes. Recent large-scale genome-wide association studies (GWAS) have identified that cis-miR-eQTLs have been prevalent near the regulatory single nucleotide polymorphisms (SNPs). However, the relationship between targeting miRNAs and cis-miRNAs of nearby genes in the context of a cooperative and regulatory function have not been elucidated. In this study, we developed a bioinformatics pipeline that can process targeting, cis-, and their associated expression information of input miRNAs using public databases (TarBase and miRmine) to report the biological significance for them. Specifically, the initial miRNA dataset was taken from a published study under the criteria of enrichment for cis-miR-eQTLs and having a false discovery rate (FDR) of less than 0.1. Under these cutoff criteria, we obtained 18 cis-miRNAs for further analysis. We filtered variants associated with the cis-miRNAs based on SNP functional categories and obtained 43 variants located in the 3' untranslated region. We employed the chromosome and variant coordinates to obtain 28 genes containing the above reported variants located in the 3' UTR using the UCSC genome browser. We searched TarBase to identify targeting miRNAs for 28 genes; out of the 28 genes, 20 had targeting miRNAs reported. Our pipeline includes a module which can preprocess a publicly available miRNA expression database - miRmine to calculate the highest expression values and their associated tissues for all known miRNAs. We then searched preprocessed expression data for 260 miRNAs targeting 20 genes and also for 18 cis-miRNAs, to retrieve the highest expression values and associated tissues. We also checked for overlaps between targeting and cis-miRNA lists. Our pipeline reported 5 common miRNAs highly expressed in the brain (hsa-miR-130b-3p), plasma (hsa-miR-196b-5p, hsa-miR-22-3p, hsa-miR-26b-5p), and blood (hsa-miR-941). To examine the functional significance for overlapping miRNAs, we ran miRmut2GO for the above identified 5 miRNAs and reported that there are 97 functional enriched categories for hsa-miR-941.
Nr: 26
Title:

Measurement of Phase Shift of Alternating Current (AC) / Voltage to Characterize Some Properties of Blood
Authors:

Nadia M. Antonova, Roumen K. Zlatev, Rogelio A. Ramos and Margarita Stoytcheva
Abstract: The blood properties reflecting the organism health condition can be determined applying biological, chemical, physical (optical, electrical) and mechanical methods. Electrical conductivity and impedance are the only electrical parameters used so far for this purpose in combination with some rheological such as blood viscosity, fractional volume concentration of red blood cells or hematocrit and others. In this work a LabVIEW platform based virtual instrument was developed for measurement of the AC current/voltage phase shift caused by a blood sample at 100 mV p-p AC voltage application within the frequency range between 1 Hz and 10 KHz. The capacitance and the active resistance values of the blood sample determining the AC current/voltage phase shift depend on blood sample properties such as blood cells adsorption and orientation as well as on their hematocrit. Some diseases strongly affect the first two parameters and hence the phase shift value. The combination of the AC current/voltage phase shift with the rheological parameters mostly the blood viscosity, allows to improve the understanding and interpretation of hemorheological results in terms of blood circulation, which could be of great medical diagnostic importance. Acknowledgements: The study was supported by the Bulgarian National Science Fund - the basic research project (№ КП-06-Н27/13 from 2018): “Development of experimental microfluidic system and methodology for assessing microrheological properties of blood. Analysis of the peripheral vasomotor reactivity and vascular endothelial function in patients with type 2 diabetes mellitus”
Nr: 29
Title:

Development of Experimental Microfluidic Device and Methodology for Assessing Microrheological Properties of Blood
Authors:

Nadia M. Antonova, Khristo Khristov, Alexandrova Anika, Muravyov Alexei and Velcheva Irena
Abstract: Background: The microfluidics has become a prominent field for studying blood microrheology. The aim of the study is to develop an experimental microfluidic device for assessing microrheological properties of blood cells’ suspensions. Methods: A method for assessing the deformability of blood cells and a device for its implementation developed by one of the co-authors was used. Based on it a new microfluidic device was elaborated and connected in a system, including a microscope with a digital camera, a pump with a manometer, computer with specially developed software. Diluted blood cells’ suspensions are investigated between two parallel optical slides with a 100 μm distance between them. The motion of the blood cells in the microchamber is observed by the microscope and recorded and visualized by the digital camera. Results: The system was tested with model suspensions and with diluted red blood cell (RBC) suspensions to specify and validate the experimental conditions: pressure range, flow rate, etc. The pressure changes, realized by the syringe pump, connected to a manometer are established and thus the changes of the shear rate in the microfluidic device are determined. Experimental data about the blood microrheological parameters as RBC aggregation and deformability, leukocyte adhesion, as well as the simultaneous evaluation of RBC aggregation and leukocyte adhesion for a group of healthy donors (control group) and from donors with type 2 diabetes (T2DM) and their mutual interaction are obtained. Conclusions: The developed device and experimental system is a promising tool for the study of erythrocyte deformability and aggregation, as well as leukocyte adhesion.
Nr: 33
Title:

Tissue Regenerator: A Device for Controlled Tissue (re-)Generation and Quality Monitoring
Authors:

Paul Ritter, Christian Lesko, Julian Bauer, Jana Dietrich, Oliver Friedrich and Michael Haug
Abstract: Whole organ and tissue engineering of donor material with the goal of revitalization and regrowth using recipient cell material are one of the utmost challenges in modern biomedical research and personalized medicine. Grown organs, be it a top-down de- and recellularization scheme, or a bottom-up 3D-printed or otherwise manufactured bioscaffold, need an optimized long-term environment with standardized process control measures to ensure reseeded tissue growth with minimized risk for organ rejection. We aim to pioneer current tissue engineering bioprocesses via complex process control, gentle stem cell inoculation, and precise (bio-) chemical treatment concepts to create an optimal tissue growth environment. During the regrowth process, our so called Tissue Regenerator measures passive parameters like biomechanics (i) e.g. scaffold compliance through resting length-tension curves, geometrical parameters (ii) 3D volume reconstructions to measure scaffold morphologies, physical characteristics (iii), e.g. changes in CIELAB color metrics and UV-VIS spectrum, vascular pressure (iv) for perfusion recellularization, and other ‘classic’ environmental parameters (v) like pO2, pH and temperature. To ensure an optimal maturation process, mechanical and electrical (direct or field) stimulation is applied depending on the organ of interest. An organ is deemed ready for transplantation via a fuzzy logic control scheme and uses time as a variable and not as a determinant. The system operates cloud-based and is optimized for remote monitoring access. To validate the system, we decellularized an M. gastrocnemius medialis via perfusion of the main feeding arteries. We showed that passive elasticity can be implemented as an additional online decellularization quality measure, noting a threefold stiffness decrease in acellular constructs. Our vision is to produce personalized organs with few risks of organ rejection.
Nr: 38
Title:

Development of an Application for Interactive Exploration and Quantification of Diagnostic Patterns of Myopathy in Muscular Tissue Histology Images
Authors:

Jacopo Baldacci, Marco Calderisi, Anna Rubegni and Filippo M. Santorelli
Abstract: Muscle biopsy is a routine diagnostic procedure, used to investigate the causes of muscle diseases. Alongside the clinical examination, electrodiagnostic, laboratory and molecular genetic testing, muscle biopsy has a critical role, providing diagnostic evidence that either establishes a disease etiology or focuses the differential diagnosis. Routine histochemistry, typically performed on frozen tissue, commonly includes various stains, which allow the assessment of muscle fiber morphology, and identification of many pathological and oftentimes diagnostic patterns. Unfortunately, the lack of uniformity in the interpretation of these patterns by clinicians is an important issue in the management of neuromuscular disorders. In fact, some pathological patterns are impossible to quantify with the naked eye presenting the risk of subjective interpretation by clinicians, which could lead to a lack of diagnosis. In this study, we develop a software that allows for an interactive exploration of two important diagnostic patterns: increased fiber size variation and increase in the number of internal nuclei. These patterns are typically visible, at light microscope, in muscle biopsy of patients affected by myopathy. Our software gets as input all the images that compose the whole scanning of the muscle section, acquired with the Zeiss Piezo-driven XY scanner. A segmentation algorithm is performed on each image to recognize the edges and separate all the fibers from each other. A machine learning algorithm based on random forest, pre-trained on a dataset of features and properties extracted from x fibers, is used to select and remove bad segmented fibers from the analysis. The software can therefore calculate the area of each segmented fiber and plot the distribution of the areas’ size. These quantifiable pieces of information could be useful for the clinician to collect objective data regarding fiber size variability and to detect important signs of myopathic damage. Moreover, we develop an algorithm of blob detection, based on the difference of Gaussian, which is able to detect internal nuclei in each fiber. Our software allows to calculate the percentage of fibers with internal nuclei and to count how many internal nuclei each fiber contains. Hence, clinicians can gather numeric and objective information about the pathological increase in the number of internal nuclei that cannot be spotted with the naked eye. This kind of data is useful to detect the extent of myopathic damage. In conclusion, the software we develop could be useful for clinicians to detect two of the most important patterns of myopathic damage and to get numeric and quantifiable data about these patterns. Our software aims to make the diagnosis independent from the interpretation of each single clinician and to contribute to the achievement of agreed and shared criteria, which are often missing especially in the diagnosis of rare disorders.
Nr: 41
Title:

Is It Worth to Model Age Impact on Functionality Status of T Cell Receptor Genes in Subgroups of young and Older Donors Separately?
Authors:

Justyna Mika, Serge Candéias and Joanna Polanska
Abstract: T lymphocytes play an essential role in the defense against pathogens and cancers through their clonally distributed T cell receptor (TR). TR genes are assembled from discrete V, D and J segments in developing lymphocytes. Due to the random nature of this process, only one-third of the rearranged TR genes are functional while 2/3 are non-functional. Here we intend to determine the impact of age on the proportion of functionally rearranged TR genes in blood by creating different age subgroups for men and women separately. We used a collection of 587 human TRβ repertoires obtained from healthy donors. After data preprocessing, we calculated diversity of functional status of TR genes (further denoted as Status Diversity) for each donor using Pielou’s J index. It ranges from 0 to 1 indicating extreme and even proportions of functional and nonfunctional genes respectively. Next, we applied a linear regression model to estimate the age relations with Status Diversity for men and women separately. The strength of relationship was estimated by Pearson correlation coefficient. Using piecewise linear regression we distinguished different patterns of evolution in age categories, using brute force approach and Bayesian Information Criterion (BIC) for model selection. The scope coefficients of the models were compared using 95% confidence intervals. Correlation coefficients between age and Status Diversity were compared between age subgroups using Cohen’s q effect size. We observe a significant decline in Status Diversity with age only in women (p=0.014, slope coefficient b1=-6.3e-04±0.0005), whereas in men the observed decline is not meaningful (p=0.454, b1=-1.9e-04±0.0005). With piecewise regression, the best split of age was found between 19 and 20 years of age in men and between 7 and 8 years of age in women. Comparison of slope coefficients showed a significant difference in the rate of diversity decline between young (b1=-7.2e-02±0.0026) and older (b1=-2.3e-05±0.0007) men. In the case of young women, slope coefficient (b1=-1.3e-02±0.0138) was not statistically significant and was not different from the one in older women (b1=-5.6e-04±0.0006). The comparison of correlation between age and Status Diversity showed that it is highly different between the young and older subsets of donors for men and women (q=0.84 and q=0.99 respectively). Altogether these results show that it is worth to analyze age relationships with Status Diversity for age categories in men, where we observe a steep decrease of diversity in young age and next stabilization over time. In women, on the other hand, there is a slight but even decrease of Status Diversity throughout the whole age range, suggesting that their peripheral T cell homeostatic expansion is regular during life. The work was supported by European Social Fund grant POWR.03.02.00-00-I029 and SUT’s grant for Support and Development of Research Potential.
Nr: 43
Title:

Exploring the Systemic Microvascular Response to a Suprasystolic Limb Occlusion: A Pilot Study
Authors:

Henrique Silva and Nicole Lavrador
Abstract: Suprasystolic limb occlusion (SLO) is a standard challenge test applied in medical research to evaluate the endothelial function of large vessels as well as of the microvasculature, allowing the assessment of cardiovascular risk. Despite the longevity of its use, several aspects of the physiological response to SLO are still currently under discussion, namely the nature of the systemic vascular response it evokes. We aimed to explore whether a SLO of the upper limb evoked a systemic microvascular adaptation in healthy subjects. Ten young healthy volunteers (21.6 +/- 2.1 years old, both sexes, 5 males) participated in this study after giving informed written consent. Following a 20 min acclimatization period to the room conditions (22-24°C) subjects performed a standard SLO protocol on a randomly chosen upper limb, divided in three phases as follows: 10 min resting (baseline), 5 min arm occlusion at 200 mmHg (challenge), 10 min post-occlusion (recovery). Local blood flow was assessed by reflectance photoplethysmography on the distal phalanx of the index fingers; skin temperature was assessed with an negative temperature coefficient thermistor placed on the middle phalanx of the index fingers; electrodermal activity (EDA) was assessed with a pair of electrodes attached to the distal phalanx of the middle and ring fingers. All signals were acquired in both limbs, with one serving as test and the contralateral as control. Nonparametric statistics were applied and a p<0.05 value was adopted. Occlusion reduced local blood flow and skin temperature in both limbs, but only significantly in the test limb. The rate of temperature decrease in the test limb was significantly higher than in the control limb. Electrodermal activity increased significantly in both limbs, suggesting an increased activity of the sympathetic nervous system, related to the perceived discomfort of the subjects. No significant difference in the relative increase of EDA signals was detected between limbs. During recovery all signals returned to their baseline values, with no significant differences between recovery and baseline being detected. These preliminary results suggest that the SLO of the upper limb evokes a systemic sympathetic-mediated microvascular response, likely responsible for the moderate perfusion reduction in the contralateral limb.