This article provides a comprehensive survey of bias mitigation methods for achieving fairness in Machine Learning (ML) models. We collect a total of 341 publications concerning bias mitigation for ML classifiers. These methods can be distinguished based on their intervention procedure (i.e., pre-processing, in-processing, post-processing) and the technique they apply. We investigate how existing bias mitigation methods are evaluated in the literature. In particular, we consider datasets, metrics, and benchmarking. Based on the gathered insights (e.g., What is the most popular fairness metric? How many datasets are used for evaluating bias mitigation methods?), we hope to support practitioners in making informed choices when developing and evaluating new bias mitigation methods.

AimsCancer diagnostics have been evolving rapidly. In England, the new National Health Service Genomic Medicine Service (GMS) provides centralised access to genomic testing via seven regional Genomic Laboratory Hubs. The PATHways survey aimed to capture pathologists’ experience with current diagnostic pathways and opportunities for optimisation to ensure equitable and timely access to biomarker testing.MethodsA nationwide survey was conducted with consultant pathologists from regional laboratories, via direct interviews based on a structured questionnaire. Descriptive analysis of responses was undertaken using quantitative and qualitative methods.ResultsFifteen regional centres completed the survey covering a median population size of 2.5 (1.9–3.6) million (each for n=12). The median estimated turnaround time (calendar days) for standard molecular markers in melanoma, breast and lung cancers ranged from 2 to 3 days by immunohistochemistry (excluding NTRKfus in breast and lung cancers, and PD-L1 in melanoma) and 6–15 days by real-time-PCR (excluding KIT for melanoma), to 17.5–24.5 days by next-generation sequencing (excluding PIK3CA for breast cancer). Tests were mainly initiated by pathologists and oncologists. All respondents discussed the results at multidisciplinary team (MDT) meetings. The GMS roll-out was perceived to have high impact on services by 53% of respondents, citing logistical and technical issues. Enhanced education on new pathways, tissue requirements, report interpretation, providing patient information and best practice sharing was suggested for pathologists and other MDT members.ConclusionOur survey highlighted the role of regional pathology within the evolving diagnostic landscape in England. Notable recommendations included improved communication and education, active stakeholder engagement, and tackling informatics barriers.

Beta-Propeller Protein-Associated Neurodegeneration (BPAN) is one of the commonest forms of Neurodegeneration with Brain Iron Accumulation, caused by mutations in the gene encoding the autophagy-related protein, WDR45. The mechanisms linking autophagy, iron overload and neurodegeneration in BPAN are poorly understood and, as a result, there are currently no disease-modifying treatments for this progressive disorder. We have developed a patient-derived, induced pluripotent stem cell (iPSC)-based midbrain dopaminergic neuronal cell model of BPAN (3 patient, 2 age-matched controls and 2 isogenic control lines) which shows defective autophagy and aberrant gene expression in key neurodegenerative, neurodevelopmental and collagen pathways. A high content imaging-based medium-throughput blinded drug screen using the FDA-approved Prestwick library identified 5 cardiac glycosides that both corrected disease-related defective autophagosome formation and restored BPAN-specific gene expression profiles. Our findings have clear translational potential and emphasise the utility of iPSC-based modelling in elucidating disease pathophysiology and identifying targeted therapeutics for early-onset monogenic disorders.

ABSTRACT This paper presents the results of cognitive interviews with 8-year-old children from four European countries – Croatia, France, Finland, and Ireland. The aim of the interviews was to pre-test a selection of well-being-related questions as a part of questionnaire development for the first European multinational birth cohort study – Growing up in Digital Europe (GUIDE)/EuroCohort. Unlike most previous studies, we focused on a younger and more age-homogenous sample, as well as a more diverse set of well-known questionnaires. A total of 68 children participated in the study. The main suggestion for the interviewing procedure is to create a safe environment yet minimize the parents’ interference in answering. The questionnaires should use child-friendly vocabulary, tangible examples, avoid complex sentence structure and negative statements. The use of timeframes in questions should be minimal. The children can use Likert-type scales, but the number of different scales in the questionnaire should be limited.

The generation of entanglement in mixed states is relevant to quantum systems coupled to an environment. The dissipative and mixing properties of the environment are unavoidable in physical platforms for quantum simulation and information processing, where entanglement can be a vital resource. In this work, we explore entanglement and heterogeneity in random Lindbladian dynamics describing open quantum systems. We propose a model of a one-dimensional chain of noninteracting, spinless fermions coupled to a local ensemble of baths. The jump operator mediating the interaction with the bath linked to each site has a power-law tail with an exponent p. We show that the system undergoes volume-to-area law entanglement phase transition in the mixed steady state by tuning p which remains stable in the presence of coherent hopping. Unlike the entanglement transition in the pure-state quantum trajectories of open systems, this transition is exhibited by the averaged steady-state density matrix of the Lindbladian. The steady state in the area-law phase is characterized by a spatial heterogeneity in local population imbalance, while the jump operators exhibit a constant participation ratio of the sites they affect. Our work provides a theoretical description of an entanglement transition realized in long-ranged open quantum systems and provides an avenue to stabilize quantum correlations in mixed states. Published by the American Physical Society 2024

AbstractMethyltransferases are gaining traction as a method to achieve greener and regioselective methylations of a wide array of substrates. In this work, a halide methyltransferase biomethylation cascade was successfully adapted with the incorporation of a methionine adenosyltransferase to generate the S‐adenosyl‐L‐methionine (SAM) methylation cofactor in situ from less costly ATP and L‐methionine, instead of directly adding SAM to start the cascade. Furthermore, this cascade was applied to achieve novel catechol and tetrahydroisoquinoline methylations with the O‐methyltransferases RnCOMT, MxSafC and NpN4OMT with high conversions and regioselectivities. In addition, the cascade was successfully trialled with the less toxic methyl group donor, methyl tosylate.

By embedding a spatially explicit ecosystem services modelling tool within a policy simulator we examine the insights that natural capital analysis can bring to the design of policies for nature recovery. Our study is illustrated through a case example of policies incentivising the establishment of new natural habitat in England. We find that a policy mirroring the current practice of offering payments per hectare of habitat creation fails to break even, delivering less value in improved flows of ecosystem services than public money spent and only 26% of that which is theoretically achievable. Using optimization methods, we discover that progressively more efficient outcomes are delivered by policies that optimally price activities (34%), quantities of environmental change (55%) and ecosystem service value flows (81%). Further, we show that additionally attaining targets for unmonetized ecosystem services (in our case, biodiversity) demands trade-offs in delivery of monetized services. For some policy instruments it is not even possible to achieve the targets. Finally, we establish that extending policy instruments to offer payments for unmonetized services delivers target-achieving and value-maximizing policy designs. Our findings reveal that policy design is of first-order importance in determining the efficiency and efficacy of programmes pursuing nature recovery. This article is part of the theme issue 'Bringing nature into decision-making'.

In recent years, there has been an increasing recognition that when machine learning (ML) algorithms are used to automate decisions, they may mistreat individuals or groups, with legal, ethical, or economic implications. Recommender systems are prominent examples of these ML systems that aid users in making decisions. The majority of past literature research on recommender systems fairness treats user and item fairness concerns independently, ignoring the fact that recommender systems function in a two-sided marketplace. In this article, we propose CP-FairRank , an optimization-based re-ranking algorithm that seamlessly integrates fairness constraints from both the consumer and producer side in a joint objective framework. The framework is generalizable and may take into account varied fairness settings based on group segmentation, recommendation model selection, and domain, which is one of its key characteristics. For instance, we demonstrate that the system may jointly increase consumer and producer fairness when (un)protected consumer groups are defined on the basis of their activity level and main-streamness , while producer groups are defined according to their popularity level. For empirical validation, through large-scale on eight datasets and four mainstream collaborative filtering recommendation models, we demonstrate that our proposed strategy is able to improve both consumer and producer fairness without compromising or very little overall recommendation quality, demonstrating the role algorithms may play in avoiding data biases. Our results on different group segmentation also indicate that the amount of improvement can vary and is dependent on group segmentation, indicating that the amount of bias produced and how much the algorithm can improve it depend on the protected group definition, a factor that, to our knowledge, has not been examined in great depth in previous studies but rather is highlighted by the results discovered in this study.

Patient and public involvement and engagement (PPIE) in the design of trials is important, as participant experience critically impacts delivery. The Edmond J Safra Accelerating Clinical Trials in PD (EJS ACT-PD) initiative is a UK consortium designing a platform trial for disease modifying therapies in PD. The integration of PPIE in all aspects of trial design and its evaluation throughout the project. PwP and care partners were recruited to a PPIE working group (WG) via UK Parkinson's charities, investigator patient groups and participants of a Delphi study on trial design. They are supported by charity representatives, trial delivery experts, researchers and core project team members. PPIE is fully embedded within the consortium's five other WGs and steering group. The group's terms of reference, processes for effective working and PPIE evaluation were co-developed with PPIE contributors. 11 PwP and 4 care partners have supported the PPIE WG and contributed to the development of processes for effective working. A mixed methods research-in-action study is ongoing to evaluate PPIE within the consortium. This includes the Patient Engagement in Research Scale -a quantitative PPIE quality measure; semi-structured interviews -identifying areas for improvement and overall impressions of involvement; process fidelity- recording adherence; project documentation review - identifying impact of PPIE on project outputs. We provide a practical example of PPIE in complex projects. Evaluating feasibility, experiences and impact of PPIE involvement in EJS ACT-PD will inform similar programs on effective strategies. This will help enable future patient-centered research.

The conformational preferences of N-((6-methylpyridin-2-yl)carbamothioyl)benzamide were studied in solution, the gas phase and the solid state via a combination of NMR, density functional theory (DFT) and single crystal X-ray techniques. This acyl thiourea derivative can adopt two classes of low energy conformation, each stabilized by a different 6-membered intramolecular hydrogen bond (IHB) pseudoring. Analysis in different solvents revealed that the conformational preference of this molecule is polarity dependent, with increasingly polar environments yielding a higher proportion of the minor conformer containing an NH⋅⋅⋅N IHB. The calculated barrier to interconversion is consistent with dynamic behaviour at room temperature, despite the propensity of 6-membered IHB pseudorings to be static. This work demonstrates that introducing competitive IHB pathways can render static IHBs more dynamic and that such systems could have potential as chameleons in drug design.