ELECTRONIC THESAURUS ON TRANSLATION STUDIES AS DATA MINING TOOL

У статті представлено результати експерименту зі впорядкування електронного тезаурусу перекладознавчих термінів, що підготовлено у рамках практичних занять з перекладної комп’ютерної лексикографії і термінографії. Тезаурус поповнюється термінами з галузі перекладознавства українською, іспанською і португальською мовами. Потреба в упорядкуванні тезарусу продиктована значною невпорядкованістю перекладознавчої термінології, відсутністю дефініції для використовуваних термінів або ж суттєвими розбіжностями між наявними дефініціями. Тезарус дозволяє добувати такі парадигматичні характеристики терміна: гіперонім, гіпонім, меронім, голонім, причина, наслідок, характеризація та інші. З-поміж синтагматичних характеристик терміна у тезаурусі надаються прикметникові і дієслівні колокації. Перші можуть пролити світло на якісні характеристики явища/предмета, позначуваного терміном; другі здатні охарактеризувати дії, виконувані предметом або ж дії, впливу яких зазнає цей предмет. Отже, семантична парадигма терміна ґрунтується переважно на його зв’язках з іншими термінами; частково парадигматичні відношення можуть бути виведені з його синтагматичних характеристик: прикметникові колокації терміна закономірно вказують на його гіпоніми. Характеризація терміна за вказаними характеристиками (зокрема, гіперонім, а також інші суттєві семи: причина, наслідок, меронім, голонім та інші) можуть бути використані для розробки дефініції терміна за інтенсіоналом. На підставі зв’язків між тезарурсними статтями можуть генеруватися нові знання, зокрема, завдяки транзитивності більшості тезаурусних функцій можуть автоматично створюватися ланцюги тезаурусних зв’язків; завдяки інверсійному пошуку можна знаходити не лише статті, в якій вихідний термін зазначено у лемі, а й його вживання у глосі й відповідні семантичні зв’язки з іншими термінами.

Creating Brand Image Profile by Social Media Analysis

The purpose of creating a brand image profile is to measure the brand perception of consumers considering brand attributes. Thus, marketing decisions can be made based on the brand's strengths and weaknesses by determining them. The brand image profile is traditionally created using the attitude scales and surveys. However, alternative methods are needed since the questionnaires' responses are careless, the number of participants is relatively low and the cost per participant is high. In this study, as an alternative method, creating a brand image profile by analyzing social media data with artificial intelligence was made for the iPhone product. Firstly, the focus group study determined the attributes related to the last version of the iPhone. Then, between December 17th, 2019 and March 23rd, 2020, 87.227 tweets that include these attributes in English were collected from the Twitter social media platform through the RapidMiner data mining tool. Sentiment analysis was performed on collected tweets by the MeaningCloud text mining tool. In this analysis, positive and negative emotions were tried to be detected through artificial intelligence algorithms. Net Brand Reputation Score (NBR) was calculated using the positive and negative tweets amount for each attribute separately. Brand image profile was created by skew analysis using NBR values. As a result, it is thought that social media analysis can be a complementary method that can be used with traditional methods in creating a brand image profile. So, it is seen as an inevitable method to use in further studies to make sentiment analysis by processing raw data received from the Social Media platforms through artificial intelligence algorithms to transform the product label or the perspectives of an event into meaningful information.

OP218 Searching Preprint Repositories For COVID-19 Therapeutics Using A Semi-Automated Text-Mining Tool

IntroductionThe COVID-19 pandemic led to a significant surge in clinical research activities in the search for effective and safe treatments. Attempting to disseminate early findings from clinical trials in a bid to accelerate patient access to promising treatments, a rise in the use of preprint repositories was observed. In the UK, NIHR Innovation Observatory (NIHRIO) provided primary horizon-scanning intelligence on global trials to a multi-agency initiative on COVID-19 therapeutics. This intelligence included signals from preliminary results to support the selection, prioritisation and access to promising medicines.MethodsA semi-automated text mining tool in Python3 used trial IDs (identifiers) of ongoing and completed studies selected from major clinical trial registries according to pre-determined criteria. Two sources, BioRxiv and MedRxiv are searched using the IDs as search criteria. Weekly, the tool automatically searches, de-duplicates, excludes reviews, and extracts title, authors, publication date, URL and DOI. The output produced is verified by two reviewers that manually screen and exclude studies that do not report results.ResultsA total of 36,771 publications were uploaded to BioRxiv and MedRxiv between March 3 and November 9 2020. Approximately 20–30 COVID-19 preprints per week were pre-selected by the tool. After manual screening and selection, a total of 123 preprints reporting clinical trial preliminary results were included. Additionally, 50 preprints that presented results of other study types on new vaccines and repurposed medicines for COVID-19 were also reported.ConclusionsUsing text mining for identification of clinical trial preliminary results proved an efficient approach to deal with the great volume of information. Semi-automation of searching increased efficiency allowing the reviewers to focus on relevant papers. More consistency in reporting of trial IDs would support automation. A comparison of accuracy of the tool on screening titles/abstract or full papers may help to support further refinement and increase efficiency gains.This project is funded by the NIHR [(HSRIC-2016-10009)/Innovation Observatory]. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

Fibromine is a multi-omics database and mining tool for target discovery in pulmonary fibrosis

Idiopathic pulmonary fibrosis is a lethal lung fibroproliferative disease with limited therapeutic options. Differential expression profiling of affected sites has been instrumental for involved pathogenetic mechanisms dissection and therapeutic targets discovery. However, there have been limited efforts to comparatively analyse/mine the numerous related publicly available datasets, to fully exploit their potential on the validation/creation of novel research hypotheses. In this context and towards that goal, we present Fibromine, an integrated database and exploration environment comprising of consistently re-analysed, manually curated transcriptomic and proteomic pulmonary fibrosis datasets covering a wide range of experimental designs in both patients and animal models. Fibromine can be accessed via an R Shiny application (http://www.fibromine.com/Fibromine) which offers dynamic data exploration and real-time integration functionalities. Moreover, we introduce a novel benchmarking system based on transcriptomic datasets underlying characteristics, resulting to dataset accreditation aiming to aid the user on dataset selection. Cell specificity of gene expression can be visualised and/or explored in several scRNA-seq datasets, in an effort to link legacy data with this cutting-edge methodology and paving the way to their integration. Several use case examples are presented, that, importantly, can be reproduced on-the-fly by a non-specialist user, the primary target and potential user of this endeavour.

SicknessMiner: a deep-learning-driven text-mining tool to abridge disease-disease associations

Background Blood cancers (BCs) are responsible for over 720 K yearly deaths worldwide. Their prevalence and mortality-rate uphold the relevance of research related to BCs. Despite the availability of different resources establishing Disease-Disease Associations (DDAs), the knowledge is scattered and not accessible in a straightforward way to the scientific community. Here, we propose SicknessMiner, a biomedical Text-Mining (TM) approach towards the centralization of DDAs. Our methodology encompasses Named Entity Recognition (NER) and Named Entity Normalization (NEN) steps, and the DDAs retrieved were compared to the DisGeNET resource for qualitative and quantitative comparison. Results We obtained the DDAs via co-mention using our SicknessMiner or gene- or variant-disease similarity on DisGeNET. SicknessMiner was able to retrieve around 92% of the DisGeNET results and nearly 15% of the SicknessMiner results were specific to our pipeline. Conclusions SicknessMiner is a valuable tool to extract disease-disease relationship from RAW input corpus.

Treosulfan-induced myalgia in pediatric hematopoietic stem cell transplantation identified by an electronic health record text mining tool

Treosulfan is increasingly used as myeloablative agent in conditioning regimen prior to allogeneic hematopoietic stem cell transplantation (HSCT). In our pediatric HSCT program, myalgia was regularly observed after treosulfan-based conditioning, which is a relatively unknown side effect. Using a natural language processing and text-mining tool (CDC), we investigated whether treosulfan compared with busulfan was associated with an increased risk of myalgia. Furthermore, among treosulfan users, we studied the characteristics of given treatment of myalgia, and studied prognostic factors for developing myalgia during treosulfan use. Electronic Health Records (EHRs) until 28 days after HSCT were screened using the CDC for myalgia and 22 synonyms. Time to myalgia, location of pain, duration, severity and drug treatment were collected. Pain severity was classified according to the WHO pain relief ladder. Logistic regression was performed to assess prognostic factors. 114 patients received treosulfan and 92 busulfan. Myalgia was reported in 37 patients; 34 patients in the treosulfan group and 3 patients in the busulfan group (p = 0.01). In the treosulfan group, median time to myalgia was 7 days (0–12) and median duration of pain was 19 days (4–73). 44% of patients needed strong acting opiates and adjuvant medicines (e.g. ketamine). Hemoglobinopathy was a significant risk factor, as compared to other underlying diseases (OR 7.16 95% CI 2.09–30.03, p = 0.003). Myalgia appears to be a common adverse effect of treosulfan in pediatric HSCT, especially in hemoglobinopathy. Using the CDC, EHRs were easily screened to detect this previously unknown side effect, proving the effectiveness of the tool. Recognition of treosulfan-induced myalgia is important for adequate pain management strategies and thereby for improving the quality of hospital stay.

Prophage Tracer: precisely tracing prophages in prokaryotic genomes using overlapping split-read alignment

The life cycle of temperate phages includes a lysogenic cycle stage when the phage integrates into the host genome and becomes a prophage. However, the identification of prophages that are highly divergent from known phages remains challenging. In this study, by taking advantage of the lysis-lysogeny switch of temperate phages, we designed Prophage Tracer, a tool for recognizing active prophages in prokaryotic genomes using short-read sequencing data, independent of phage gene similarity searching. Prophage Tracer uses the criterion of overlapping split-read alignment to recognize discriminative reads that contain bacterial (attB) and phage (attP) att sites representing prophage excision signals. Performance testing showed that Prophage Tracer could predict known prophages with precise boundaries, as well as novel prophages. Two novel prophages, dsDNA and ssDNA, encoding highly divergent major capsid proteins, were identified in coral-associated bacteria. Prophage Tracer is a reliable data mining tool for the identification of novel temperate phages and mobile genetic elements. The code for the Prophage Tracer is publicly available at https://github.com/WangLab-SCSIO/Prophage_Tracer.

THE DEEP TIME OF BITCOIN: EXCAVATING THE "WORK" IN PROOF-OF-WORK CRYPTOCURRENCY SYSTEMS

Estimates place Bitcoin’s current energy consumption at 141.83 terawatt-hours/year, an amount comparable to Ukraine. While Bitcoin’s energy problem has become increasingly visible in both academic and popular discourse (see Lally et al. 2019), the computational mechanisms through which the Bitcoin network generates coins, proof-of-work, has gone under-examined. This paper interrogates the “work” in proof-of-work systems. What is this work? How can we access its material history? I trace this history through a media archaeology of computational heat, in an attempt to better situate the intimate relationship between information and energy in proof-of-work systems. I argue the “work” in these systems is principally heat-work, and trace its ideological constructions back to nineteenth-century thermodynamic science, and the reframing of doing work as something exhaustible, directional, and irreversible (Prigogine & Stengers 2017; Daggett 2019). I then follow thermodynamic discourse through Cybernetics debates in the 1940s, illustrating how, early in the formation of Information Theory, the heat-work undergirding the functioning of a “bit” was obscured and compartmentalized, allowing information to be productively abstracted apart from its energetic infrastructures (Hayles 1999; Kline 2015). I conclude with a discussion of the heat-work within the Application Specific Integrated Circuit (ASIC), Bitcoin’s principal mining tool, arguing that proof-of-work mining is not a radical exception to the computing status quo, but rather a lens through which to think more broadly about computing’s complex relationship to energy, and ultimately, how this relationship can be different.