143174 research outputs found
Sort by
Tailoring butane-1,2,3,4-tetraol-based maltosides (BTMs) via group-swapping and detergent unsymmetry: new detergent design strategies for membrane protein studies
No full textMembrane proteins are essential bio-macromolecules involved in numerous critical biological processes and serve as therapeutic targets for a wide range of modern pharmaceuticals. Small amphipathic molecules, called detergents or surfactants, are widely used for the isolation and structural characterization of these proteins. A key requirement for such studies is their ability to maintain membrane protein stability in aqueous solution, a task where conventional detergents often fall short. While many new detergents have been developed based on novel molecular scaffolds, comparatively little effort has been made to enhance detergent performance through rational modification of existing structures, largely due to the limited availability of guiding design principles and strategies. In this study, we refined previously developed butane-1,2,3,4-tetraol based maltosides (BTMs), using two structural modification strategies, head/tail group-swapping and the introduction of hydrophobic unsymmetry. The resulting group-swapped (GS)-BTMs exhibited distinctive physical properties compared to the original BTM, including differences in water-solubility (∼7 to >10 wt%), critical aggregation concentration (5 to 15 μM), and self-assembly size (7.6 to 34.2 nm). When evaluated using model membrane proteins, including the human adrenergic receptor (β2AR), symmetric GS-BTMs (e.g., GS-BTM-C11 and GS-BTM-C12) showed superior performance relative to the original BTM-C11 and benchmark detergents (DDM and LMNG). The unsymmetric variants, such as GS-BTM-C14,10 and GS-BTM-C15,9, further improved protein stability. These findings highlight group-swapping and hydrophobic unsymmetry as effective strategies for enhancing detergent performance. This work demonstrates how minimal structural modifications can impact detergent properties and efficacy, providing valuable insights for the development of improved detergents from existing molecular frameworks
Mapping use of high dose or long-term oral glucocorticoids and steroid-sparing strategies in adults with chronic conditions: a rapid scoping review of reviews
No full textPurpose: Oral glucocorticoids (OGCs) have a broad range of uses and are effective in treating numerous conditions. However, it is commonly acknowledged that OGCs at high doses or over long periods have a burden of toxicity. Despite the use of steroid-sparing therapies, OGCs continue to be prescribed to treat a wide range of immune and inflammatory conditions. We aimed to address the following research questions: 1) what are the contemporary indications for high dose and/or long-term OGCs? 2) what patterns of use are described in the literature for high dose and/or long-term OGCs? 3) which evidence do we have for chronic conditions related to well-established steroid-sparing strategies and tapering regimes for OGCs? 4) what adverse effects have been reported with high dose and/or long-term OGCs?
Methods: A rapid scoping review was conducted using the Joanna Briggs Institute guidelines. The Protocol has been published on the Open Science Framework. A systematic search of MEDLINE (Sep 2014 to Sep 2024) identified systematic reviews and scoping reviews involving adults (≥18 years) treated with high dose and/or long-term OGCs for chronic inflammatory conditions. Studies involving pregnant women were excluded.
Results: In total 137 reviews were included. OGCs were indicated in 47 different conditions in dermatology, respiratory, gastrointestinal, haematology, immunology, respiratory, rheumatology and other miscellaneous categories. Across all specialties, OGCs were used either at high doses (at least 20 mg prednisone equivalent per day) or for long durations (for at least 3 months). For types of adverse effects reported in the included reviews, 20 were labelled as endocrine, 13 as immunological, 21 as musculoskeletal, 30 as gastrointestinal, and 16 as cardiovascular. Sixty-four reviews looked for/reported unspecified adverse events. One hundred and fifteen reviews had evidence of steroid-sparing/tapering regimes, indicating the wide use of these strategies to mitigate the harmful effects of OGCs.
Conclusion: OGCs are used for a broad range of inflammatory conditions across multiple specialties. There is evidence related to a broad range of potential adverse effects across multiple body systems regardless of indication of use. Further research is needed using a combined cross-condition approach to their measurement and reduction, alongside gaining more insight into the impact of OGCs on patients’ quality of life
The cost of dengue shock and septic shock in Vietnam: a patient-centred economic analysis
No full textBackground
Dengue shock (DS) and septic shock (SS) are the most common infectious causes of shock in Vietnam. Little is known about the cost of an episode of DS or SS from the patient perspective. We aimed to describe the direct medical, non-medical and productivity costs associated with DS and SS.
Methods
We recruited adults with DS and SS to a prospective observational study at the Hospital for Tropical Diseases, Ho Chi Minh City, from 2019 to 2021. We collected hospital bills, insurance status and out of pocket payments, and conducted an economic questionnaire at discharge, 1, 3 and 6 mo later. We calculated the proportion incurring catastrophic health expenditure (CHE) and catastrophic costs.
Results
We recruited 127 adults with DS, of whom 118 survived, and 35 with SS, of whom 24 survived; 18.9% and 71.4% with DS and SS, respectively, incurred CHE. When non-medical and productivity costs were considered, the true cost of illness was 6.3 and 6.7 times higher than the hospital bill for DS and SS, respectively.
Conclusions
Productivity costs must be counted when assessing the cost of critical illness in low- and lower-middle income countries. It is vital that financial protection systems are extended to cover patients requiring high-cost critical care in Vietnam
Offline reinforcement learning: in pursuit of perfect policies from imperfect data
No full textIntelligent agents can solve tasks in dynamic environments without the need for human supervision. Many environments are so large and complex that specifying exactly how an agent should behave is challenging. Instead, an agent must learn through trial and error from environment feedback; this learning paradigm is reinforcement learning (RL). Rather than being explicitly told how to behave, RL agents improve their policy by interacting with the environment and learning which actions maximize the long-term utility of a reward function in a sequential decision-making process. The sample efficiency of an algorithm describes how performance scales with data. Maximizing efficiency is critical as environments grow; with it, the data needed grows exponentially due to the curse of dimensionality.
An alternative RL paradigm explores how to learn without any direct environment interaction. Given historical data of (imperfect) demonstrators attempting to complete a task, offline RL aims to recover an optimal policy in a setting that requires maximal sample efficiency.
This thesis presents three novel algorithms for offline RL. First, we design an algorithm that expands on a minimalist approach to scale to multimodal datasets without the tuning complexity that recent methods increasingly assume. Second, we extend an existing model-free method to derive an effective and efficient offline model-based RL algorithm. Finally, we reframe offline policy improvement as a regression problem that aligns the policy towards high-reward, in-sample behaviors. We evaluate our algorithm in off-policy and on-policy settings and show that our method can enjoy the stability benefits of on-policy learning while remaining competitive with both our off-policy variant and other baselines.Open Acces
Remote consultations in England during COVID-19: challenges in data quality, linkage, and research validity
No full textThe COVID-19 pandemic accelerated the adoption of remote consultations across health care, requiring rapid adjustments in service delivery. Consequently, there is an urgent need to understand the impact of remote consultations on health pathways. This viewpoint paper explores key challenges in data sources in England that hinder research on the impact of remote consultations on health outcomes. Based on our experience conducting research on this topic, we present variations in observational study findings and their validity, considering differences in population characteristics and data sources. We provide recommendations to enhance data quality for future research, including improvements in data recording platforms and strengthened structures for linking primary and secondary care electronic health records
Beyond markov decision processes: How to leverage structure to build efficient reinforcement learning algorithms
No full textA Markov Decision Process (MDP) describes a general framework for modelling decision-making in an uncertain environment. Over the years, researchers have mainly focused on developing Reinforcement Learning (RL) algorithms to learn how to behave optimally in unknown environments without making explicit assumptions about the existence of underlying structures in the MDP.
This general approach has been instrumental to RL's widespread success across various domains. However, when additional structure is present and exploitable, enabling algorithms to leverage it might lead to significant efficiency improvement. This work investigates the benefits of designing RL algorithms that can efficiently leverage these structures.
Specifically, this work focuses on two distinct types of known latent structures.
First, we consider MDPs that exhibit a hierarchical structure; that is, tasks described by such MDPs can be decomposed into a sequence of sub-tasks. In this context, we provide a lower bound on the sample complexity of hierarchical RL algorithms, which allows us to quantify the potential benefit of hierarchical approaches. We also offer a framework for building hierarchical algorithms that leverage a known hierarchical decomposition. The validity of that framework is supported by theoretical guarantees of its efficiency and empirical evidence that it outperforms its monolithic counterpart whenever a hierarchical structure is present.
Second, we consider MDPs that exhibit a graphical structure. The algorithm has access to a graph that encodes the conditional independence between state variables, and the unknown dynamics can be inferred only by local components of the graph. In this context, we provide a posterior sampling-based algorithm that theoretically and empirically outperforms RL algorithms that do not leverage this structural property. Finally, we provide empirical evidence that this latent graphical structure is present in optimising wind farms' yields and demonstrate the efficiency of our algorithm on that particular task.Open Acces
Retired contact sports athletes with cognitive concerns: promoting lifelong brain health
No full textThere is widespread concern amongst former athletes about the link between head injury and dementia. Neurologists are increasingly assessing ex-contact sports athletes with cognitive and behavioural issues following repetitive head impacts and traumatic brain injury. Their assessment and management can be challenging due to the broad differential diagnosis, including psychiatric issues, trauma-related impairment and, in some cases, neurodegeneration. There may be a range of pathologies present after trauma exposure, including Alzheimer’s disease and chronic traumatic encephalopathy. Currently, we have only limited understanding of specific clinical phenotypes for distinct types of post-traumatic dementia, nor are there in vivo tests for many of the pathologies. Informed by our experience running a midlife brain health clinic for retired elite contact sport athletes, we describe a practical framework for the workup of athletes with cognitive concerns, highlighting key clinical features, an approach to investigation including neuroimaging and advanced fluid biomarkers, symptomatic management strategies, and research directions
Mixed data regression for price prediction
No full textThere are many applications of price regression: to manage cash flows for business, balance supply and demand for energy, or quantify risks for insurance.
Historically, regression models -- be they linear or complex neural networks -- have mainly used `tabular' data composed of numeric or nominal categorical variables.
However, as companies and institutes increasingly collect more text data, and language models substantially improve every year, there is growing interest in how to leverage text for regression.
This thesis studies using text-only data as well as combining text and tabular data for regression, applied to price prediction.
We begin by discussing an often overlooked aspect of tabular modelling: missing value handling.
A large, real UK insurer dataset is used to study the impact of missing value handling on the performance of gradient boosted trees and deep learning models, two contemporary tabular modelling approaches.
In doing so we establish Catboost as a strong tabular model, relevant for benchmarking joint text-tabular models, but emphasize the importance of missing values.
We then conduct a study of text-only regression.
Text-only regression requires a pipeline of preprocessing, tokenization, featurization and modelling steps.
We rigorously analyze choices for these steps, where the literature lacks systematic work.
Among other results, for the first time, our study quantifies the relative importance of the model choice compared to choice of steps before modelling.
We find that choice of the upstream steps contributes variance in performance comparable to model choice.
We conclude with a study of joint text-tabular modelling: studying an ensemble of a tabular-only Catboost and a text-only neural network.
This ensemble is compared to existing methods.
Overall, we conclude that although this ensemble does not consistently have the best accuracy, it is simple, robust to data with low signal in the text modality and achieves good accuracy.Open Acces
Value of clinical review for AI-guided deep vein thrombosis diagnosis with ultrasound imaging by non-expert operators
No full textDeep vein thrombosis (DVT) carries high morbidity, mortality, and costs globally. Point of care ultrasound (POCUS) image acquisition by non-ultrasound-trained providers, supported by an AI-based guidance and remote image review system, is believed to improve the timeliness and cost-effectiveness of diagnosis. We examine a database of 381 patients with suspected DVT who underwent an AI-guided ultrasound scan by a non-ultrasound-trained nurse and an expert sonographer-performed standard compression ultrasound scan. Each AI-guided scan was reviewed remotely by blinded radiologists or blinded independent POCUS-certified American Emergency Medicine (EM) physicians. Remote reviewer and standard scan diagnoses were compared. The primary endpoint is AI-guidance system sensitivity with clinician review; secondary endpoints include specificity, positive predictive value, negative predictive value, image quality, inter-observer image quality, and vein compressibility agreement. Data was analysed through the bootstrapping method, bootstrapping with a second reader for each scan, and a majority voting system. Eighty percent (n = 304) of scans were of sufficient diagnostic quality. Radiologist reviewer sensitivity ranged from 90%–95%, specificity from 74%–84%, NPV from 98%–99%, PPV from 30%–42%, and potential expert-led ultrasound scans avoided from 39%–50%. Inter-observer agreement for image quality was 0.15 and for compressibility 0.61. EM reviewer sensitivity ranged from 95%–98%, specificity from 97%–100%, NPV was 99%, PPV from 81%–100%, and potential expert-led ultrasound scans avoided from 29%–38%. Inter-observer agreement for image quality was 0.59 and for compressibility 0.67. Diagnosing lower extremity DVT through AI-guided image acquisition with clinician review is feasible. Performance is influenced by reviewer expertise. We find potential positive impacts on health economics, including safely avoiding expert-led ultrasound scans
Near-Unity Nitrate to Ammonia conversion via reactant enrichment at the solid-liquid interface
No full textElectroreduction of nitrate (NO3‒) to ammonia (NH3) is a promising approach for addressing energy challenges. However, the activity is limited by NO3‒ mass transfer, particularly at reduction potential, where an abundance of electrons on the cathode surface repels NO3‒ from the inner Helmholtz plane (IHP). This constraint becomes pronounced as NO3‒ concentration decreases, impeding practical applications in the conversion of NO3‒-to-NH3. Herein, we propose a generic strategy of catalyst bandstructure engineering for the enrichment of negatively charged ions through solid-liquid (S-L) junction-mediated charge rearrangement within IHP. Specifically, during NO3‒ reduction, the formation of S-L junction induces hole transfer from Ag-doped MoS2 (Ag-MoS2) to electrode/electrolyte interface, triggering abundant positive charges on the IHP to attract NO3‒. Thus, Ag-MoS2 exhibits a ~ 28.6-fold NO3‒ concentration in the IHP than the counterpart without junction, and achieves near-100% NH3 Faradaic efficiency with an NH3 yield rate of ~20 mg h‒1 cm‒2 under ultralow NO3‒ concentrations