A5013970475- Multiple Sclerosis Research Studies
- Machine Learning in Healthcare
- Advanced Graph Neural Networks
- Neural Networks and Applications
- SARS-CoV-2 and COVID-19 Research
- Gaussian Processes and Bayesian Inference
- Time Series Analysis and Forecasting
- Model Reduction and Neural Networks
- Computational Drug Discovery Methods
- Machine Learning in Materials Science
- Digital Imaging for Blood Diseases
- AI in cancer detection
- Single-cell and spatial transcriptomics
- Artificial Intelligence in Healthcare
- Metabolomics and Mass Spectrometry Studies
- Mycobacterium research and diagnosis
- Medical Image Segmentation Techniques
- Multiple Myeloma Research and Treatments
- Complex Systems and Time Series Analysis
- Statistical Methods in Clinical Trials
- COVID-19 epidemiological studies
- Cancer-related molecular mechanisms research
- Advanced Causal Inference Techniques
- Bioinformatics and Genomic Networks
- Topological and Geometric Data Analysis
Yale University
2023-2025
KU Leuven
2019-2024
Hasselt University
2023
The University of Melbourne
2021-2022
University of Tasmania
2021-2022
University of Technology Sydney
2021
University of Amsterdam
1995
<h3>Background and Objectives</h3> People with multiple sclerosis (MS) are a vulnerable group for severe coronavirus disease 2019 (COVID-19), particularly those taking immunosuppressive disease-modifying therapies (DMTs). We examined the characteristics of COVID-19 severity in an international sample people MS. <h3>Methods</h3> Data from 12 data sources 28 countries were aggregated (sources could include patients 1–12 countries). Demographic (age, sex), clinical (MS phenotype, disability),...
Naïve CD8 T cells have the potential to differentiate into a spectrum of functional states during an immune response. How these developmental decisions are made and what mechanisms exist suppress differentiation toward alternative fates remains unclear. We employed in vivo CRISPR-Cas9–based perturbation sequencing assess role ~40 transcription factors (TFs) epigenetic modulators cell fate decisions. Unexpectedly, we found that knockout TF Klf2 resulted aberrant exhausted-like acute...
Modeling real-world multidimensional time series can be particularly challenging when these are sporadically observed (i.e., sampling is irregular both in and across dimensions)-such as the case of clinical patient data. To address challenges, we propose (1) a continuous-time version Gated Recurrent Unit, building upon recent Neural Ordinary Differential Equations (Chen et al., 2018), (2) Bayesian update network that processes sporadic observations. We bring two ideas together our...
Certain demographic and clinical characteristics, including the use of some disease-modifying therapies (DMTs), are associated with severe acute respiratory syndrome coronavirus 2 infection severity in people multiple sclerosis (MS). Comprehensive exploration these relationships large international samples is needed.Clinician-reported demographic/clinical data from 27 countries were aggregated into a set 5,648 patients suspected/confirmed disease 2019 (COVID-19). COVID-19 outcomes...
Background: We need high-quality data to assess the determinants for COVID-19 severity in people with MS (PwMS). Several studies have recently emerged but there is great benefit aligning collection efforts at a global scale. Objectives: Our mission scale-up and provide community data-driven insights as soon possible. Methods: Numerous stakeholders were brought together. Small dedicated interdisciplinary task forces created speed-up formulation of study design work plan. First step was agree...
Abstract On March 11, 2020, the World Health Organization declared COVID-19 outbreak, originally started in China, a global pandemic. Since then, outbreak has indeed spread across all continents, threatening public health of numerous countries. Although Case Fatality Rate (CFR) is relatively low when optimal level healthcare granted to patients, high percentage severe cases developing pneumonia and thus requiring respiratory support worryingly high, could lead rapid saturation Intensive Care...
Abstract Background People with multiple sclerosis (MS) are a vulnerable group for severe COVID-19, particularly those taking immunosuppressive disease-modifying therapies (DMTs). We examined the characteristics of COVID-19 severity in an international sample people MS. Methods Data from 12 data-sources 28 countries were aggregated. Demographic and clinical covariates queried, alongside outcomes, hospitalisation, admission to ICU, requiring artificial ventilation, death. Characteristics...
Disability progression is a key milestone in the disease evolution of people with multiple sclerosis (PwMS). Prediction models probability disability have not yet reached level trust needed to be adopted clinic. A common benchmark assess model development also currently lacking.
Multiple sclerosis (MS) is a chronic autoimmune disease of the central nervous system that results in varying degrees functional impairment. Conventional tools, such as Expanded Disability Status Scale (EDSS), lack sensitivity to subtle changes progression. Radiomics offers quantitative imaging approach address this limitation. This study used machine learning (ML) and radiomics features derived from T2-weighted Fluid-Attenuated Inversion Recovery (FLAIR) magnetic resonance images (MRI)...
Abstract Increasing the number of organ donations after circulatory death (DCD) has been identified as one most important ways addressing ongoing shortage. While recent technological advances in transplantation have increased their success rate, a substantial challenge increasing DCD resides uncertainty regarding timing cardiac terminal extubation, impacting risk prolonged ischemic injury, and negatively affecting post-transplant outcomes. In this study, we trained externally validated an...
<h2>Abstract</h2> The Multiple Sclerosis Data Alliance (MSDA), a global multi-stakeholder collaboration, is working to accelerate research insights for innovative care and treatment people with multiple sclerosis (MS) through better use of real-world data (RWD). Despite the increasing reliance on RWD, challenges limitations complicate generation, collection, these data. MSDA aims tackle sociological technical arising scaling up specifically focused MS envisions patient-centred ecosystem in...
Graph neural networks (GNNs) are a powerful architecture for tackling graph learning tasks, yet have been shown to be oblivious eminent substructures such as cycles. We present TOGL, novel layer that incorporates global topological information of using persistent homology. TOGL can easily integrated into any type GNN and is strictly more expressive (in terms the Weisfeiler--Lehman isomorphism test) than message-passing GNNs. Augmenting GNNs with leads improved predictive performance node...
A simple and interpretable way to learn a dynamical system from data is interpolate its vector-field with kernel. In particular, this strategy highly efficient (both in terms of accuracy complexity) when the kernel data-adapted using Kernel Flows (KF) (Owhadi Yoo, 2019) (which uses gradient-based optimization based on premise that good if there no significant loss half used for interpolation). Despite previous successes, (based interpolating vector field driving system) breaks down observed...
Background Investigating low-prevalence diseases such as multiple sclerosis is challenging because of the rather small number individuals affected by this disease and scattering real-world data across numerous sources. These obstacles impair integration, standardization, analysis, which negatively impact generation significant meaningful clinical evidence. Objective This study aims to present a comprehensive, research question–agnostic, multistakeholder-driven end-to-end analysis pipeline...
Background The integrity and reliability of clinical research outcomes rely heavily on access to vast amounts data. However, the fragmented distribution these data across multiple institutions, along with ethical regulatory barriers, presents significant challenges accessing relevant While federated learning offers a promising solution leverage insights from sets, its adoption faces hurdles due implementation complexities, scalability issues, inclusivity challenges. Objective This paper...
Abstract Multiple myeloma management requires a balance between maximizing survival, minimizing adverse events to therapy, and monitoring disease progression. While previous work has proposed data-driven models for individual tasks, these approaches fail provide holistic view of patient’s state, limiting their utility assist physician decision-making. To address this limitation, we developed transformer-based machine learning model that jointly (1) predicts progression-free survival (PFS),...
Abstract Background People with multiple sclerosis (MS) are a vulnerable group for severe COVID-19, particularly those taking immunosuppressive disease-modifying therapies (DMTs). We examined the characteristics of COVID-19 severity in an international sample people MS. Methods Data from 12 data-sources 28 countries were aggregated (sources could include patients 1-12 countries). Demographic (age, sex), clinical (MS phenotype, disability), and DMT (untreated, alemtuzumab, cladribine,...
A simple and interpretable way to learn a dynamical system from data is interpolate its vector-field with kernel. In particular, this strategy highly efficient (both in terms of accuracy complexity) when the kernel data-adapted using Kernel Flows (KF)\cite{Owhadi19} (which uses gradient-based optimization based on premise that good if there no significant loss half used for interpolation). Despite previous successes, (based interpolating vector field driving system) breaks down observed time...
Diffusion-based manifold learning methods have proven useful in representation and dimensionality reduction of modern high dimensional, throughput, noisy datasets. Such datasets are especially present fields like biology physics. While it is thought that these preserve underlying structure data by a proxy for geodesic distances, no specific theoretical links been established. Here, we establish such link via results Riemannian geometry explicitly connecting heat diffusion to distances. In...
We present a generative approach to classify scarcely observed longitudinal patient trajectories. The available time series are represented as tensors and factorized using deep recurrent neural networks. learned factors represent the data in compact way can then be used downstream classification task. For more robustness accuracy predictions, we an ensemble of those models mimic Bayesian posterior sampling. illustrate performance our architecture on intensive-care case study in-hospital...