Orange Data Mining

We are helping crops like tomatoes, peppers, and cucumbers grow healthier and stronger! 🍅 🫑 💪 Recently, Neža Pajek Arambašič and Tomaž Curk from our team at the University of Ljubljana, Faculty of Computer and Information Science travelled to the 8th Croatian Congress of Microbiology in Trakošćan, Croatia 🇭🇷 , to present their research in a session on AI for Microbiology. 📌 Neža gave a presentation titled: Uncovering hidden genetic diversity in the genus Tobamovirus through large-scale mining of public sequence repositories. 👏 Tobamoviruses are a group of plant viruses that can cause diseases in crops and other plants, interfering with how plants grow. For instance, some tobamoviruses can cause: 🌶️ stunted plant growth 🌶️ reduced plant health and yield 🌶️ damaged or lower-quality fruit. Discovering previously unrecognised viruses in publicly available data can help scientists better understand and monitor the viruses, and develop measures to limit their spread 🔝 Neža showed how, with the help of #machinelearning, scientists can search through huge amounts of publicly available genetic sequencing data (i.e., information about the genetic material of an organism, obtained by reading its genetic sequence) and identify previously unrecognised plant viruses that deserve closer expert investigation. 🧬 💻 🔍 👀 🥬 The researchers used Serratus PalmID to query 14.8M public datasets, then assembled candidates using a custom Snakemake pipeline. 🥬 Because manual curation of assembled contigs (a longer DNA or RNA sequence assembled from many smaller sequencing fragments) is a major bottleneck, they implemented a two-stage machine learning model to filter background noise and prioritise candidate contigs for expert review. 🥬 Neža's team identified more than 32 putative (suspected, but not yet confirmed) novel tobamoviruses, including new strains from Amazonian freshwater and European municipal wastewater. 🙌 This approach shows how combining #datamining, automated workflows, and machine-learning-assisted curation can enhance early virus discovery for agriculture. 💯 In the Tobamo project, our team collaborated with the National Institute of Biology 🇸🇮 More about the Tobamo project 👉 https://lnkd.in/dA6H-JP9 About the congress 👉 https://lnkd.in/dMbg4q35 #bioinformatics #AI #dataanalytics #virology #microbiology

We’re celebrating again! Bravo, Yun! 🏆 🎉 😊 We're so proud and happy to share that Yun Wang from our team at the University of Ljubljana, Faculty of Computer and Information Science, recently won the prestigious BioASQ Challenge Award. 🙌 #BioASQ is an international challenge where teams build and test #AI systems that search biomedical literature, identify relevant evidence, and use it to answer expert-written biomedical questions. 🧬 💻 🤔 🏆 Yun won the BioASQ Challenge Award for his performance in task 14b: Biomedical Semantic QA: Given a biomedical question, competing advanced AI systems need to find relevant research articles and passages, provide a precise answer, and, where appropriate, produce a short summary of the answer. 🌍 In Phase B of this task, 3️⃣ 1️⃣ 9️⃣ system entries from Asia, Canada, Europe, and the USA were submitted and evaluated across 4 batches. The training/development set contained 5,729 biomedical questions. Yun won with the system: A Multistage Evidence Retrieval System for BioASQ Task 14b: Hybrid Retrieval, Reranking, and Snippet Selection. Sincere congratulations, Yun!! 👏 💯 🎉 More about the competition 👉 https://bioasq.org/ #dataanalytics #datascience #machinelearning #medicine #healthcare #innovation

OrangeEDU is here! 🍊💜🇪🇺 Discover a DALI4US data literacy ecosystem, including: 🍯 a browser-based data analysis workflow environment, 🍯 guided interactive lessons, and 🍯 accompanying learning scenarios (with printable materials, videos, and teacher guidance)   A central objective of DALI4US is to equip primary school teachers with the confidence, knowledge, and skills required to teach #dataliteracy effectively. That includes making sense of real-world problems through data 🍂 , collecting and organising information, identifying and interpreting patterns, reasoning from evidence, and communicating findings. 📊 🤔 💬 This also provides a basis for understanding how data are used within #AI and #machinelearning systems. 🎯 One of the project’s key outputs is OrangeEDU. The technological starting point was Orange Data Mining, open-source visual #dataanalysis and machine learning software developed at the University of Ljubljana, Faculty of Computer and Information Science 🇸🇮, and used in research, higher education, and data science training. 🌍 To test both Orange and the DALI4US pedagogical approaches for developing data literacy in the classroom, the project carried out a sequence of teacher workshops. Participants told us that Orange helped them understand the data concepts being introduced. However, many found the software too complex to use independently in their own classrooms. 🫠 As teacher feedback accumulated, it became clear what adaptations were required: 🍊 developing OrangeEDU as a browser-based environment with a smaller selection of functions relevant to the learning activities, 🍊 guided online lesson pages in which selected OrangeEDU functions could be prepared in advance and embedded directly into a learning activity. We listened, and the technology and the curriculum-linked learning scenarios were developed in parallel and tested together. 👌 We are happy to announce the result, tested and evaluated in 3 countries 🇮🇪 🇱🇺 🇸🇮: OrangeEDU integrated in ready-to-use learning scenarios and classroom resources. 🙌 Ready to try it out with your students? You can access the interactive lesson plans here 👉 https://gairdin.eu/ #dataanalytics #education #innovation #EU Revelo.ai

Listen to your heart ❤️ 💯 How can we better understand the data generated during heart treatment procedures to develop more advanced and more personalised treatment methods❓ In a medical procedure called cardiac ablation, abnormal heart rhythms (arrhythmias) are treated by selectively changing a tiny area of heart tissue. 👀 🩺 A very small area of heart tissue that is causing or sustaining the abnormal rhythm is destroyed or disabled so that it no longer conducts the problematic electrical signal. ❤️🩹 In a newer type of cardiac ablation called pulsed field ablation (PFA), instead of using heat or extreme cold to damage the target tissue, very short, high-voltage electrical pulses are delivered to the heart. ⚡As electroporation can be relatively selective for cardiac muscle cells, it has the potential to reduce damage to some surrounding structures compared with conventional thermal ablation. 🔝 💻 In the CardioEP project, Nika Krivic from our team at the University of Ljubljana, Faculty of Computer and Information Science is developing a web app that makes it easier to explore, understand, and compare complex ECG recordings. Recently, she has been working on a research pipeline for analysing (anonymised) electrophysiological data acquired during cardiac ablation procedures, including PFA, from the Claris EP system at the University Medical Centre Ljubljana. 🇸🇮 Nika and the team developed: ❤️ a tutorial that walks you through the structure of the data, including loading procedures, sessions, annotations, metadata, and ECG signals, ❤️ an application that allows us to explore database statistics and individual procedures, including timelines, channels, ablation events, annotations, and the ECG signals. The current pipeline allows us to: ❤️ Extract 12-lead ECG segments from procedures, place them in the context of the relevant stage of the procedure, process them appropriately, and save them in a format suitable for further analysis, ❤️ Pass the prepared ECG segments through existing ECG foundation models to obtain embeddings, ❤️ Explore embeddings and batch effects: using different methods, we can investigate, including visually, whether observed differences represent actual clinical patterns or so-called batch effects. Next step❓ Accounting for differences in the data that may come from the way the recordings were collected. Once we reduce these technical differences, we can use #AI-generated representations of the ECGs (embeddings) to identify general patterns in healthy and pathological ECGs associated with the PFA procedure. 🔝 Follow the project here 👉 https://www.cardioep.vip/ #dataanalytics #machinelearning #medicine #healthcare #cardiology #innovation

🍊⚕️🍬 Orange friends, especially students! If you are interested in how new technologies are changing #healthcare and how to apply your knowledge to real-world challenges, don't miss this opportunity! ⤵️ Personalised medicine is bringing together #biology, data, technology, and #business to develop new approaches to understand, diagnose, and treat disease. 🧬 This transformation is creating entirely new fields of work, innovation, and entrepreneurship 👀 Explore the field through two free InnovPrecMed 🇪🇺 online courses and discover how scientific ideas can grow into clinical solutions, innovative products, and new business opportunities. 🔝 🧪 Course: From Lab to Impact 👉 https://lnkd.in/exV79yXZ Explore concrete examples of collaboration with companies such as BioVendor Group, Lightly Technologies, KP Therapeutics, Pancrevo, and MendelFOLD, and see how scientific discoveries can be transformed into breakthrough clinical and business solutions. 🧪 Course: Business Plan 👉 https://lnkd.in/eCb_5Sfm How can you turn an innovative idea into a business opportunity? Learn about the business planning process – from the fundamentals of entrepreneurship and market analysis to financial planning, risk management, and effective business plan presentation. 🍯 And there’s more! After completing the course, you can apply for a free mentoring programme 👉 https://lnkd.in/dKg9Q9nX or gain practical experience through internships with partner organisations involved in the InnovPrecMed project 👉 https://lnkd.in/dth3fGUf Upon completing the course, participants receive an official certificate of completion. 🎓 Let's go! 💯 More about the courses 👉 https://lnkd.in/eHu5Dhsn #education #innovation #dataanalytics #medicine #bioinformatics #EU

We have a new doctor in the lab! 🔥🔥🔥 Pavlin Poličar from our team at the University of Ljubljana, Faculty of Computer and Information Science, Laboratory for Bioinformatics, recently successfully defended his PhD thesis. 👏 His research focused on methods for dimensionality reduction and visual #dataanalysis, including ✨ work on openTSNE, ✨ visual explanations of embeddings, and ✨ methods for analysing changing or temporal data: Modern scientific inquiry is increasingly characterised by the analysis of massive, high-dimensional datasets. In many domains, low-dimensional embeddings have become an important tool for exploring such data: they reveal structure, support hypothesis generation, and facilitate scientific communication. 💯 However, their practical use raises methodological challenges related to scalability, reuse, interpretation, and the incorporation of temporal structure. 🤔 Pavlin’s dissertation makes three sets of contributions:   🎯 It develops scalable and reusable embedding infrastructure through openTSNE, a modular Python implementation of t-SNE 🎯 It introduces automated methods for interpreting embeddings 🎯 It develops methods for temporally aware embedding construction   🎓 Evaluated on nationwide electronic health record and single-cell transcriptomics data, these contributions show how low-dimensional embeddings can be constructed, reused, interpreted, and extended to support reasoning about complex dynamic systems.   Massive congratulations, Dr. Poličar!! 🙌 🎉 🥂   Check out Pavlin’s PhD presentation here 👉 https://lnkd.in/d3M4cnqC #dataanalytics #machinelearning #innovation #education

Orange and Quasar meet TotalEnergies in Southern France! 🍊☀️ 🇫🇷 🌍 Global companies shaping the energy of today and tomorrow are investing in #machinelearning, #dataanalytics, and #ai education. Total Energies employees from research departments joined Marko Toplak from our team at the University of Ljubljana, Faculty of Computer and Information Science, and Ferenc Borondics from Synchrotron SOLEIL at a 3-day practical workshop on Quasar. ⚡🌊 🌬️🌤️ ✈️ 🚘 Marko and Ferenc travelled to the Total Energies offices in beautiful Lacq 🌳🌻 to share their expertise in spectroscopy – a scientific method for studying objects and materials based on the characteristic patterns of light they absorb, emit, or reflect at different wavelengths (colours). Companies like Total Energies work with complex scientific and engineering data, using #datascience, AI, modelling, and simulation to address challenges across energy and the energy transition. Quasar can support this research process by making it easier to explore and visualise data, combine different datasets, and test machine-learning methods through interactive visual workflows. 💻 ✨ 💯 As an open-source extension of #OrangeDataMining, Quasar provides researchers with a flexible environment for exploring data and developing analytical approaches before moving towards more specialised or production-ready solutions. 🔝 Are you interested in a Quasar training tailored to your company needs? Get in touch 😊 Explore Quasar 👉 https://quasar.codes/ #orangeforbusiness #education #business

Have you heard about Quasar❓🍊☀️😯 We often describe it as Orange’s cousin 😁 - a variant of Orange with pre-installed tools for scientific #dataanalysis, particularly for spectroscopy, which studies what materials are made of by measuring how they interact with light. ✨ ✈️ 🌊 Recently, Marko Toplak from our team at the University of Ljubljana, Faculty of Computer and Information Science travelled to Oxford, UK 🇬🇧, to hold a 3-day hands-on InfraRed Data and Image Analysis training on Quasar for its current, prospective, and potential international users at Diamond Light Source. 🌍 💻🔝 The training was a collaboration between Diamond Light Source (MIRIAM beamline B22), Synchrotron SOLEIL (SMIS beamline), and the University of Ljubljana. The goal behind this and similar events is to promote the Quasar open-source software among the international scientific research community. Are you joining the next workshop sessions? Stay tuned. 😊 Find out more about Quasar 👉 https://quasar.codes/ #dataanalytics #machinelearning #AI #education #orangedatamining #orangeforbusiness

Science to the people ✌ The science news platform Clear Sky Science recently featured a paper co-authored by Lena Trnovec from our team! 🧡 The aim is to make science more accessible through jargon-light explanations 🙌 With her co-authors, Lena from our team at the University of Ljubljana, Faculty of Computer and Information Science explores how a soil-dwelling amoeba, Dictyostelium discoideum, keeps thousands of its cells marching through development in sync. 🦠 🥁🚶♂️➡️ Understanding this natural "cell choreography" helps explain how tissues form correctly - and what might go wrong when timing falls apart. 🧩 🧩 🕳️ 🧬 ⏱️ By showing that single-cell RNA sequencing can quantify synchronicity over time, this work also provides a blueprint for probing how timing is controlled in more complex organisms, and what could happen when that timing breaks down. 🤔 Clear Sky curates summaries of peer-reviewed work, written for curious non-specialists. 🔝 You can read the featured article here 👉 https://lnkd.in/eFAUzEBR #datascience #biology #bioinformatics #dataanalytics #machinelearning #AI #innovation #education

What can we learn from talking to fish❓🐠 💦 In the DANIO-ReCODE research project 🇪🇺, we use AI to explore how zebrafish - animals with a remarkable ability to regenerate complex organs like the brain, heart, and eyes. 🧠 💙 👀 Combining biological research with #AI, computational methods, and data visualisation to study the biological processes behind tissue regeneration, we aim to better understand how damaged tissues and organs might eventually be repaired in humans. 🤔💡 The project also provides advanced doctoral training to a new generation of early-career researchers working in the field of vertebrate tissue regeneration. 🔝 Martina Sevo from our team at the University of Ljubljana, Faculty of Computer and Information Science is one of them - contributing to the project by developing an AI-powered data chatbot! 💬 💻 🐠 In spring, Martina and other doctoral students involved in the project participated in a training at Imperial College London 🇬🇧, focused on: 🐠 the DANIO-CODE Data Coordination Center ➡️ the project's central infrastructure for organising, managing, and sharing the biological data generated within the project, and supporting consistent data processing and analysis; 🐠 omics data analysis workflows using nf-core and Nextflow ➡️ computational workflows for processing and analysing large-scale molecular datasets, with a focus on making analyses more systematic, standardised and reproducible. The engaging, practically-oriented training was led by Damir Baranasic from Ruđer Bošković Institute and Matthias Hörtenhuber from SciLifeLab Data Centre. Soon after, the DANIO-ReCODE project team met again in Vienna 🇦🇹, where Martina presented the progress in her research work: Conversational AI for Exploring Complex Data in Tissue Regeneration 👏 You can find out more about the project here 👉 https://lnkd.in/e3jvPSRq #dataanalytics #innovation #datascience #machinelearning #biology DANIO-ReCODE MSCA Doctoral Network