A. K. Virdi

Abstract Background: Fabry Disease (FD) is a rare metabolic condition, characterised by a wide range of symptoms and progressive deterioration of the heart, kidneys and nervous system. The condition often goes undiagnosed for between 3-10 years, with misdiagnoses during this period being common. A historically common misdiagnosis for FD is Hypertrophic Cardiomyopathy (HCM). This condition mimics the cardiac symptoms of FD, although it does not share the same genetic pathology. Machine learning (ML) is a subfield of artificial intelligence that has shown promise in identifying rare diseases, including FD. This study builds on previous research to identify FD in routine clinical test data and provide a robust explainability framework. Methods: Cardiac test data (ECG, Holter and Echo) was extracted for 100 HCM and 49 FD patients. Stratified repeated K-fold cross-validation procedure was employed to select the optimum ML classifier. Following this, we explored a range of ML models and ensembling approaches. Each tuned classifier was trained on this data to identify differences between the experimental groups. Feature importance, SHAP values and partial dependence plots were generated to explain the model's decision-making process. Results: Following cross-validation, the Extreme Gradient Boosting (XGB) classifier emerged as the most reliable classification algorithm (AUC and F1 score: 0.92). T-axis, gender and age were revealed as consistently being the primary features used by the model for classification predictions. Conclusion: XGB proved to be a reliable classifier between FD and HCM for this dataset. The most important features largely aligned with established clinical approaches to differentiating phenotypes between HCM and FD. Future work should aim to compare the performance of this model against expert clinicians to compare these methods to the human state of the art.

Here, we demonstrate how chemical interface damping (CID) influences the nanoscale plasmon–exciton coupling strength.

Abstract There is a vast range of gases which get ionized, produce electron-ion pairs, on the passage of high energy charged particles. Such gases are extensively used in experiments such as LHC, Belle-II, RHIC, DAFNE, etc which produce high energy particles. Panjab University established a detector assembly and characterization laboratory dedicated to gaseous detectors such as Resistive Plate Chamber (RPC) and Gas Electron Multiplier (GEM) for the LHC experiment. Here, we present the recent work on the fabrication and characterizations of the GEM detector at Panjab University.

The Compact Muon Solenoid (CMS) is a general-purpose particle detector at the Large Hadron Collider (LHC) designed to study a wide range of particles produced in high energy collisions. The interaction of the beams with the pipe, shielding and detector supporting materials can produce neutrons, photons, electrons and positrons, forming a common background radiation field for CMS detector. A Monte-Carlo simulation is used to predict the background rate for a newly installed detector. In the forward region, the upgrade includes Gas Electron Multiplier (GEM) detectors called GE1/1. In this study, an estimate of the GE1/1 detector response to the background radiation is presented. The flux of background radiation is predicted using the FLUKA framework and the response of the detector is predicted using the GEANT4 framework. A comparison with actual GEM slice data is used as validation.

The collider experiments of the modern era produce an extreme environment of radiation fields. It becomes quite challenging to operate the detectors in such an environment as the high-radiation background complicates the particle identification. Particles produced in proton–proton (pp) collisions interact with the beam pipe, shielding and the other detector supporting materials to produce charged hadrons and neutrons along with photons, electrons and positrons. These particles interact with the surrounding material acting as a common background radiation field for the CMS detector. CMS has installed new muon detectors based on Gas Electron Multiplier (GEM) technology at the endcap station 1 $$(1.55<|\eta |<2.18)$$ , called GE1/1. In this study, an estimation of the GE1/1 detector response to background radiation is presented using the FLUKA and GEANT4 frameworks.