Single-cell RNA sequencing (scRNA-seq) has transformed transcriptome profiling at cellular resolution, yet accurate reconstruction of full-length transcripts for individual cells remains a central challenge. Emerging scRNA-seq protocols can produce reads that span entire transcripts, enabling isoform-level expression analysis. For example, Smart-seq protocols combine unique molecular identifier (UMI)-linked reads that index and stitch together multiple reads from the same molecule, with internal reads filling coverage gaps. We demonstrate that these read types exhibit markedly different biological and statistical properties in strandness, 5'/3' coverage bias, and genomic locality. Existing assemblers fail to leverage these distinctions, yielding suboptimal assembly. We developed Amaranth, a novel single-cell assembler that discriminatively models UMI and internal reads. Amaranth implements heuristics specifically designed to address the distinct biases of UMI-linked and internal reads, enabling accurate strandness assignment for internal reads, reliable splicing graph refinement, and precise transcript start site determination. We also developed Amaranth-meta, which integrates information across cells to enhance individual cell assemblies. Benchmarked on Smart-seq3 datasets from human HEK293T and mouse fibroblast cells, Amaranth outperformed other state-of-the-art assemblers in assembling individual cells and in meta-assembly. Amaranth advances isoform-level analysis in single-cell transcriptomics, facilitating detailed studies at cellular resolution. Amaranth is implemented in C++ and is freely available at https://github.com/Shao-Group/amaranth under the BSD-3-Clause license. Scripts, documentation, and data for reproducing experiments in this manuscript are available at https://github.com/Shao-Group/amaranth-test.
Stroke-induced writing disorders offer valuable insights into the neural mechanisms of writing. The Japanese writing system is particularly useful for such investigations, as it includes both kana (phonograms) and kanji (morphograms). Kana is primarily associated with phonological processing, whereas kanji relies on lexical-orthographic pathways. Although previous research suggests there are distinct neural substrates for kana and kanji, most studies have focused on small cohorts with pure agraphia, and large-scale investigations of aphasia remain underexplored. To address this gap, we examined a large stroke cohort to identify anatomical differences underlying phonogram versus morphogram processing. We analysed 315 patients with post stroke aphasia who underwent a comprehensive battery of language assessments, including kana and kanji writing tests, and MRI at multiple stroke centres between 2016 and 2024. Using multivariate support vector regression-based lesion-symptom mapping and structural disconnection analyses based on a continuous permutation-based family-wise error correction, we investigated associations between lesion/disconnectome maps and various language scores, with a particular focus on writing scores. To complement voxel-wise mapping, we prespecified a tract-level multivariable regression to quantify disconnection load across language-related white matter tracts. We evaluated 315 patients (mean age 67.2 ± 13.3, 34.6% female). Lesion-symptom mapping suggested that impairments in kana writing were associated with left frontal regions, whereas lesion clusters associated with kanji writing emerged primarily in the white matter underlying the supramarginal and angular gyri. Disconnectome analyses implicated left dorsal pathways (e.g. arcuate fasciculus, superior longitudinal fasciculus) in kana writing impairments, with left ventral pathways (e.g. inferior fronto-occipital fasciculus, inferior longitudinal fasciculus) implicated in kanji writing impairments. Complementary tract-level regression mirrored this dissociation: kana outcomes showed significant associations with dorsal pathways, whereas kanji outcomes were predominantly associated with ventral pathways. These findings highlight distinct neural pathways for writing phonograms (kana) and morphograms (kanji), providing novel insights into the neural mechanisms of writing disorders. In this large multicentre aphasia cohort, disconnectome mapping and tract-level regression provide convergent evidence for a dorsal-ventral dissociation between pathways associated with phonogram- and morphogram-related writing impairments. Our results contribute to understanding script-specific neural processing and may inform future assessment and rehabilitation strategies for aphasia in languages with complex writing systems.
Interleukin-33 (IL-33) is classically viewed as an alarmin, but in cancer, it may do more than merely initiate immunity. I propose that IL-33 scripts the trajectory of immunity through interconnected regulatory layers that influence how host responses are organized and ultimately resolved into either productive antitumor immunity or suppressive, repair-like states.
Transaction verification is essential to blockchain security. As blockchain data continue to grow, resource limited nodes may be forced to operate as non-full nodes, which weakens independent verification and may increase centralization risk. To address this issue, the stateless blockchain technology has been proposed, which uses the accumulators to combine the UTXO set into one fixed-size commitment. However, they suffer from two critical limitations: (i) the inability to support script validation duo to the lack of scriptSig, and (ii) the absence of an outsourcing mechanism to ensure that task executors reliably provide the appropriate witness for the nodes just recovered from failures. We propose LSTVS, a lightweight stateless transaction verification architecture for UTXO based blockchains, which extends RSA accumulator based stateless verification with script-based authorization verification, cache assisted stale proof tolerance, and outsourced witness updates. First, we incorporate UTXO fields associated with transactions into the membership witness to enable digital signature verification. Second, we reconstruct the transaction data format to prevent the exponential growth of transaction reference fields. Finally, we introduce an outsourcing mechanism to improve the transaction verification rate while minimizing computational resource consumption. Experimental results show that the proposed architecture supports the core validation dimensions of UTXO based stateless verification, including existence verification, unspent status verification, and script-based authorization verification, while avoiding UTXO scale dependent proof growth and introducing input dependent transient witness overhead. Compared with existing state of the art RSA accumulator-based schemes, LSTVS improves the transaction verification rate and reduces the local witness update overhead for intermittently online nodes.
Intensive care unit (ICU) nurses improve patient outcomes through respiratory weaning protocols, yet this is not standard practice in the neonatal ICU (NICU). Infants born extremely prematurely, <28 weeks gestational age, are at risk for the development of severe bronchopulmonary dysplasia (BPD). Reducing exposure to invasive mechanical ventilation (iMV) can reduce the severity of BPD, yet it remains a challenge. This quality improvement initiative assessed the effect of optimizing nurse-led rounds scripts on the interdisciplinary rounding structure and explored iMV time among infants in a BPD cohort. The 3 objectives were education, script compliance, and iMV time. The project was guided by the Plan-Do-Study-Act cycle model and Lewin's Change Theory. Implementation occurred in a level IV NICU within a single children's hospital. The population included 240 NICU nurses, 57 neonatal nurse practitioners, and 118 neonatologists caring for a cohort of 18 patients with BPD. Nurses-led rounds through an optimized respiratory script for a cohort of patients with BPD. After 90 days of implementation, nurses' confidence, autonomy, and respiratory knowledge were compared. iMV time for the patients with BPD was retrospectively reviewed. Daily script usage improved to 61% by day 90. Education was completed by 99% of the nurses. The nurse's confidence, knowledge, and sense of autonomy all increased. These results indicate that nurse-led rounds scripts benefit NICU nurses' leadership, improve interdisciplinary collaboration, and improve respiratory outcomes for NICU patients.
We introduce Actigraphic Data Analyser (ADA), a user-friendly open source Python package with a graphical user interface (GUI). It reads raw data from GENEActiv and ActiGraph devices, and the MESA dataset, supports multiple methods for collapsing data into epochs, provides sleep/wake scoring via several classical algorithms and the recently published Universal Filter Approach, offers comprehensive circadian rhythms analysis, and outputs common sleep plots and metrics like daily profiles, sleep onset, sleep fragmentation index and more. Apart from the interactive GUI, ADA can be used as a Python library. As an example of its application, we present a simple Python script revealing correlations between different descriptors of circadian rhythms: DFA, variants of cosinor, AR model spectrum, IS, IV, M10 and L5. Correlation matrix produced by this script (a figure included also in this paper) applied to a dataset of 87 weekly recordings (created for this study and freely available) reveals division of these commonly applied descriptors into two groups: one related mostly to the (length of) circadian period, and the other to the strength of the 24-h rhythm. Both the above-mentioned ADA package and set of weekly actigraphic recordings are freely available on GPLv3 and CC-BY licences, respectively.
Population genomic workflows frequently rely on fragmented command-line utilities, custom conversion scripts, and programming language-specific environments, complicating computational reproducibility and obscuring data provenance. As analytical workflows become increasingly automated and computationally intensive, dependence on disparate preprocessing tools can introduce friction between raw genotype files, quality-control decisions, statistical analyses, and downstream workflows. We developed SNPio, a Python-native framework that consolidates single nucleotide polymorphism data parsing, filtering, visualization, numerical genotype encoding, and population genomic summary-statistic calculation within a unified software architecture. VCF file parsing and filtering benchmarks were compared against vcfR and SNPfiltR. SNPio demonstrated faster execution times but used more memory than its R-based comparators, reflecting SNPio's retention of genotype arrays, metadata, and provenance-tracking attributes. Pairwise Weir and Cockerham's FST and Nei's genetic distance estimates aligned with HierFstat expectations based on Pearson correlations and aggregate error metrics. D-statistics conformed to theoretical expectations across eleven simulated datasets spanning a range of introgression signal strengths. SNPio provides a reproducible Python-native workflow for processing, filtering, encoding, visualizing, and analyzing SNP datasets. It integrates common early-stage population genomic operations into a transparent, scriptable framework, which ultimately promotes workflow provenance and reduces reliance on disjointed software tools, unsaved terminal commands, and custom scripts. SNPio is particularly suited for population genomic studies of non-model organisms in ecological, evolutionary, and conservation contexts, where reproducible preprocessing and interoperability with downstream analyses are becoming increasingly important.
<p><strong>Introduction:</strong> Multi-centre datasets are becoming essential for evidence-based decision-making in geriatric otolaryngology, yet their reliability depends on a transparent and reproducible process of data preparation.</p><p><strong>Aim:</strong> To present the complete "engine-room" workflow of the ENT Geriatric Database - a harmonised, anonymised multi-centre dataset - and to demonstrate how systematic data curation supports trustworthy analyses of perioperative risk and outcomes in elderly patients.</p><p><strong>Methods: </strong>The study integrated 16 Excel files from eight tertiary ENT centres into a unified dataset of 888 patients aged 65 years and older. The scripted pipeline standardised 97 canonical variables using positional and keyword-based mapping, removed personally identifiable information, harmonised heterogeneous coding schemes across 44 variables, and performed a multilevel data-quality audit.</p><p><strong>Results:</strong> The workflow revealed major structural and clinical data-quality issues: five variables exceeded 50% overall missingness, 44 had coded-missing categories, and 99 potential duplicates were detected. Five centre-specific block patterns indicated how procedural differences influence data completeness. All cleaned data, missingness matrices, and anomaly logs were exported for independent review.</p><p><strong>Conclusions:</strong> A scripted, auditable curation pipeline can transform heterogeneous spreadsheets into a more reliable multicentre dataset. Investing in this "engine-room" infrastructure strengthens data validity and provides a more robust basis for perioperative risk estimation and meaningful inter-centre comparisons in geriatric otolaryngology.</p>.
Advanced Persistent Threat campaigns have increasingly adopted semantic obfuscation techniques in malicious Office macros, rendering the code logic opaque to traditional scrutiny. Despite the decline in volumetric attacks following Microsoft's default blocking policy, these sophisticated vectors can bypass traditional static syntactic pattern matching and evade dynamic analysis through environment awareness guardrails. To recover logic hidden by such obfuscation, this paper proposes a static analysis framework centered on semantic-aware code reconstruction. Unlike conventional methods, our approach reconstructs the underlying execution logic from obfuscated scripts to extract hidden Indicators of Compromise (IoCs). A notable feature of this framework is the Obfuscation Awareness and Splitting Approach. This algorithmic mechanism addresses the challenges of context window limitations and logical fragmentation by utilizing quantitative metrics to identify high-density obfuscation zones and optimally partition scripts, ensuring the preservation of semantic continuity during reconstruction. We then employ a generative semantic engine to process these partitions, feeding a hybrid feature extraction pipeline for multidimensional threat characterization. We systematically evaluate the framework on a contemporary dataset of obfuscated malicious macros. Experimental results demonstrate that our semantic reconstruction approach achieves an average precision of 74.57% in IoC extraction, outperforming conventional static analysis in our evaluation. When integrated with machine learning classifiers, the framework attains a maximum detection accuracy of 98.89%. The experimental results indicate the effectiveness and robustness of our semantic deobfuscation-based framework in real-world malware detection, offering enterprises a scalable solution for defensive deployment.
This article presents BACI-VI-Bench, a processed benchmark dataset and reproducible construction pipeline that transforms CEPII-BACI product-level international trade records into finite-dimensional variational inequality instances for multi-commodity trade-network equilibrium. The source data are CEPII-BACI HS17 bilateral trade files, version V202601, covering annual exporter-importer-product flows from 2017 to 2024. The pipeline filters positive bilateral trade flows, maps six-digit Harmonized System products to commodity sectors, selects leading exporting and importing economies, aggregates values and quantities into structured flow tensors, normalizes observed trade volumes, and calibrates a Nagurney-style benchmark cost-and-price operator. The dataset provides year-level instances of dimension 500 (five commodity sectors, ten exporters, ten importers, one route) and sector-level instances of dimension 100 for five HS commodity groups: Machinery, Minerals, Chemicals, Transport Equipment, and Metals. Each variational inequality instance contains a feasible set, a calibrated operator, observed and normalized flow tensors, exporter and importer identifiers, sector labels, operator parameters, and projection residual diagnostics. The repository includes Python construction and validation scripts, benchmark characterization data for extragradient and self-adaptive inertial solvers, figures, and metadata files enabling reuse with projection, extragradient, self-adaptive variational inequality algorithms, and multi-agent reinforcement learning frameworks. The benchmark dataset was originally developed in the context of a broader study on trade-network equilibrium modeling. Nevertheless, the present data article provides a complete and self-contained description of the dataset, benchmark construction pipeline, instance format, metadata, validation procedures, and reproducibility resources, enabling independent reuse of the benchmark.
Computerized radiotherapy chart checking tools have revolutionized the initial physics chart review process as they offer verification of large amounts of plan parameters in seconds and allow physicists to concentrate on high-skill-level items that are challenging for automation. Both commercialized and in-house chart check solutions typically rely on access to the patient data within the live SQL database of the radiation oncology information system (ROS). This is a complex task potentially posing risks to both database integrity and overall system performance when accessed during clinic operation hours. The aim of this study was to develop and test a chart checking tool based on AURA reports that utilize the reporting database to detect and analyze the errors occurring during treatment plan preparation in external beam radiotherapy. An Automated AURA Report-based Chart Checking Tool (AARCCT) was developed using the python-based programming environment in the RayStation treatment planning system (TPS). Following TG-275 recommendations, the tool verifies specific physics check items in TPS plan data and Varian's ARIA ROS. The ARIA data was captured leveraging advanced Physics Summary (AURA) reports. The AARCCT was tested on > 600 patients receiving various modalities of treatment, including 3D-CRT, IMRT and electron treatments. In addition to applying the script to the plans immediately following plan development (i.e., before physics check), it was applied to ∼160 plans already reviewed by medical physicists. The detected errors were assessed with failure mode and effect analysis. The AARCCT was able to analyze over fifty plan parameters in near to real-time. Before physics review, ∼48.6% of plans contained at least one error, largely low severity. The error rate was relatively constant throughout the year of testing. After physics checks were completed, AARCCT detected errors in ∼37.1% of physicists checked plans. Both before and after physics check, the most common errors were related to inaccuracies in patient setup imaging (24%), prescription (5.6%), written directive (6.3%), patient shifts (5.4%) and contours (2.6%). The error occurrence rate across the dosimetry team was found to be between ∼22% to 62% with no correlation to dosimetrist's experience. The relative error occurrence rate across the radiation oncologist (RO) team was found to be ∼40%-60%. The higher error rates were observed in ROs who were either recent hires or who had > 20 years of experience. Across the physics team, the error occurrence rate was ∼22%-47% with higher rates among those with less than three or more than twenty years of experience. The AARCCT was found to be an essential tool to reduce error propagation following manual physics plan checks. The automated nature and omission of live ROS database access uniquely allows for smooth integration of the developed tool into the clinical workflow while minimizing the impact to ROS speed, functionality and security. The tool could also be used for evaluation of staff training and establishing uniformity of practice across the radiation therapy team members to improve quality and operational efficiency.
Cancer vaccines inducing immunogenic responses to tumor-specific neoantigens are rapidly emerging into a new frontier of cancer therapy. Chimeric RNAs encoding fusion proteins are a rich source of novel neoantigens. Here, we present a straightforward bioinformatic pipeline to identify immunogenic peptides produced from chimeric RNAs. We apply a bespoke script to identify fusion-specific peptide regions from chimeric transcript predictions and leverage the netMHCpan program to identify immunogenic peptides. In this paper, we provide a guide for installation and running of the program as well as discuss the rationale behind its design.
Psychologists have long studied people's responses to adverse life events. Certain ways of telling a story of suffering, favored by a person's culture, may be more adaptive than others. The present study explores the culturally sanctioned script for how to tell the story of the lowest point in one's life. The present study introduces the Psycho-Social Script for Suffering (PSSS), a four-step narrative sequence through which a story of suffering features (1) situational context, (2) emotional expression, (3) closure, and (4) positive meaning. From a sample of 158 Black and white American midlife adults interviewed at three time points over 9 years, the authors coded 426 low point stories for the four narrative themes and analyzed their relationship to self-report measures. Narrating low point stories with higher levels of PSSS was positively associated with measures of well-being, maturity, and adaptive personality traits. Moreover, with a significant main effect of time, PSSS increased over the 9 years as indicative of a developmental sequence in midlife. The longitudinal analysis and rigorous mixed-methods approach offer novel insights into the potential benefits of narrating suffering in a way that resonates with culturally accepted discursive norms.
To describe the quality of nurse-patient interactions when nurses use Electronic Health Record systems in four acute hospital wards. An explanatory sequential mixed methods study. Sixteen researcher observations were conducted using a published Quality of Interactions Schedule tool to evaluate the duration and quality of nurse-patient interactions. Observations were followed by 16 nurse and 16 patient interviews. Data were analysed by descriptive statistics and thematic analysis and integrated to inform overall study meta-themes. Three study meta-themes emerged: (1) Limited social, open, reciprocal and face-to-face nurse-patient communication; (2) Cumbersome computer systems monopolised nurses' time and attention and impeded face-to-face communication; and (3) Nurses' use of Electronic Health Record scripts fostered a task-orientated agenda. Nurses, healthcare employers and system developers need to consider the unintended impact of nurses' use of Electronic Health Records on the quality of nurse-patient interactions. Nurses need to evaluate practices that promote, and not hinder, quality nurse-patient interactions when nurses use Electronic Health Record systems in acute care settings. Researchers developing Electronic Health Record systems need to involve nurses and patients. Balancing the complex tripartite relationship between the nurse, patient and digital interface has implications for nursing practice, education and research. The challenges encountered when nurses use Electronic Health Records need to be addressed to promote quality nurse-patient interactions. Less obtrusive Electronic Health Record technology is required that is developed with nurses who are the principal users. Nurse educators need to promote techniques that facilitate person-centred communication when nurses use Electronic Health Record systems and researchers need to evaluate practices that promote quality nurse-patient interactions. Digital transformation will continue to dominate nursing in the future and the significant findings from this study will help inform further exploration and developments in this area. Patients consented to the collection of data and for the data to be used in future potential publications. Patient participants were all discharged from the acute care hospital soon after data collection.
The present research was designed to determine whether an ambiguous, visually presented event is better recalled if an emotional (relative to neutral) verbal interpretation of the event is read before or after seeing the video. There are two competing hypotheses. First, researchers have found that emotional events are better recalled relative to neutral events. As such, one possibility is that the presentation of an emotional verbal interpretation of the event - read before or after the video itself - may enhance subsequent memory of the event. Alternatively, research on the "verbal overshadowing effect" shows that the subsequent verbal description of an event can impair memory for the event itself. This suggests that information presented asynchronously to the video may adversely affect memory for the video. We showed participants (N = 649) 2-min videos that could be interpreted in either a mildly or a very negative emotional way. Before or after viewing a video, people were given a script that allowed for a neutral or negative verbal interpretation of the video itself, with the negative interpretation causing them to have a more robust emotional response to the video. Memories of the video were then assessed either immediately or following a 1- or 7-day delay. Memory of both the video (using detail, inference, and wrong probes) and the text (using verbatim, paraphrase, inference, and wrong probes) were examined. Results revealed an "emotional verbal overshadowing effect," such that emotional information presented asynchronously to the video produced the greatest decrement in subsequent memory.
Motivational interviewing (MI) is an effective approach for supporting health behaviorchange, but face-to-face delivery is resource-intensive and difficult to scale. Rule-based conversational agents (CAs) can improve access; however, their scripted interactions and limited language flexibility constrain MI delivery. While large language models (LLMs) are increasingly being used for MI coaching, their conversational fidelity and quality compared with human coaches and rule-based CAs remain understudied. This study aimed to describe the development of an LLM-based CA, Artificially Intelligent Motivational Interviewing (Aimi), orchestrated with structured workflows, and to evaluate its feasibility, conversational fidelity, and user perceptions during MI coaching interactions. We developed Aimi using structured LLM workflows designed to enhance MI fidelity. We conducted a within-participants study, where 18 adults interacted with (1) Aimi, (2) a novice MI-trained human coach, and (3) a rule-based CA during live text-based role-play coaching sessions. Transcripts were independently evaluated by an MI expert using the Motivational Interviewing Skill Code, Version 2.0 (MISC-2), to assess MI competency and fidelity. Participants completed a user experience questionnaire to provide general feedback and to assess session alliance, dialogue relevance, empathy, engagement, linguistic quality, and perceived motivation to change. Feedback from users was thematically summarized and categorized under strengths and weaknesses for each approach. Aimi achieved fidelity scores comparable to those of the novice human coach and higher than those of the rule-based CA on summary metrics, including higher reflection-to-question ratios (median 0.84, IQR 0.62-0.92 vs 0.62, IQR 0.42-0.74 vs 0.25, IQR 0.17-0.38), more complex reflections (median 66.67%, IQR 46.97%-76.92% vs 50%, IQR 34.38%-61.88% vs 0.00%, IQR 0%-50%), and greater elicitation of client change talk (median 90.83%, IQR 85.89%-100% vs 73.21%, IQR 63.10%-83.19% vs 66.67%, IQR 57.86%-81.94%). User experience ratings showed no significant differences across conditions. User feedback revealed distinct strengths and limitations across the coaching interactions. Participants described Aimi's interactions as personalized, fluid, and adaptive, though sometimes overly reflective and lengthy. The novice human coach was viewed as empathetic and supportive but slow to respond, whereas the rule-based coach was viewed as efficient and structured yet limited in depth and personalization. This study demonstrates the technical feasibility of structured LLM-workflows for MI coaching and their capacity to maintain conversational fidelity comparable to that of a novice MI-trained human coach. Given the role-play paradigm, single-rater coding, and small convenience sample, these comparative findings should be interpreted as exploratory. Our findings serve as a foundational baseline for the development of scalable behavior change interventions in clinical settings.
Pathway and functional enrichment analysis is a cornerstone of omics data interpretation, enabling researchers to map differentially expressed proteins or genes onto curated biological processes, signaling cascades, and molecular functions. While tools such as Ingenuity Pathway Analysis (IPA), g:Profiler, and Enrichr are widely used to generate ranked enrichment results, translating these tabular outputs into clear, publication-ready figures remains a time-consuming step that typically requires custom scripting and familiarity with visualization libraries - a significant barrier for researchers without a computational background. Here we present EnrichViz, a self-contained, browser-based R Shiny application that enables interactive, code-free visualization of pathway and functional enrichment results from quantitative proteomics, transcriptomics, and metabolomics experiments. EnrichViz accepts three standard CSV files as input - a normalized abundance matrix, a sample annotation or metadata file, and enrichment results from any platform that exports tabular output - and produces six complementary, publication-ready visualizations: bar and bubble plots for ranking enriched terms by significance, chord diagrams for exploring pathway-molecule connectivity, clustered heatmaps for displaying Z-score normalized expression patterns across experimental groups, and boxplots or violin plots for examining the abundance distribution of individual proteins, genes, or metabolites. The application supports both raw p-values and pre-transformed -log10(p) values through automatic detection, and all plot parameters are adjustable in real time through a graphical sidebar. Every figure can be exported as a high-resolution PNG file at 300 dpi. EnrichViz is implemented in R using the Shiny, ggplot2, pheatmap, and circlize packages, and is freely available at https://rgmilian.shinyapps.io/EnrichViz/ .
This study aims to analyse the transition process experienced by mother/father dyads when they recognise their children's gender dysphoria. The study employed a qualitative approach, based on interpretative description and the theoretical support of the transitions theory, in which 10 dyads participated. The interviews were conducted using a semi-structured script, recorded, transcribed, and subjected to reflective thematic analysis. Revealed a central theme: transitioning together. This entails the facilitating and inhibiting factors that best describe the experiences of these dyads. The theory is a resource for healthcare professionals to recognise phenomena that help or hinder the transition processes of children and adolescents.
The use of health systems data (HSD) such as patient records and prescriptions is expanding within clinical trials to improve efficiency, reduce participant burden, and enhance real‑world relevance. However, challenges remain around governance, data quality, transparency and public trust, echoed through discussions in this project. While Patient and Public Involvement and Engagement is increasingly embedded in trial design, involvement in methodological and data‑focused areas remains limited. Ensuring that HSD trial training for researchers and public reflects public perspectives is therefore important. This work was conducted within HDR UK's Transforming Data for Trials programme. A Participatory Health Research (PHR) approach was used to develop training resources for the HDR UK Futures online learning platform. A UK‑wide Public Advisory Group (PAG) was recruited through national involvement networks. Engagement took place through virtual full-PAG and small-working‑group meetings. PAG members were involved in identifying priorities, shaping content, reviewing materials and recording videos. Two case studies illustrate different approaches to involvement: co‑production throughout development of a public‑facing module and consultative input to a researcher‑focused module. The PAG comprised 25 members with diverse geographical and experiential backgrounds. Case Study 1: members helped develop eight short videos for public partners. They influenced topic selection, tone, language and accessibility of scripts; six members also participated in filming. Their input led to clearer explanations of topics including data governance, consent and trust, with content presented in plain English and modular formats. Case Study 2: PAG members provided feedback on a technical training module on Data Utility Comparison Studies which compare the usefulness of different data sources. Their feedback highlighted concerns about data accuracy, transparency and potential equity implications, prompting refinements to the framing and examples used within the module. Co‑production supported the development of accessible resources for public research partners, while consultative input helped ensure researcher‑focused training addressed issues relevant to public trust and accountability. These approaches demonstrate how public perspectives can strengthen training related to HSD trial methodology. Embedding public perspectives in HSD training development could enhance relevance, accessibility and trust. This work provides a practical model for involving public contributors in methodological training within trials.
Phylogenetic analyses of entire genomes (phylogenomics) have revealed abundant heterogeneity of evolutionary histories. While much has been done to model this heterogeneity and to infer species trees despite it, the current toolkit has a limitation. Most methods assume that gene trees across the genome differ but are all sampled from the same distribution, defined by models such as the multi-species coalescent (MSC), and parametrized consistently across the genome. Empirical data strongly suggest this assumption is often violated because the species tree, its parameters, or the process generating the gene trees can all change across the genome. Errors in the data can further compound this heterogeneity. To address this challenge, we define the problem of detecting what segments of the genome are inconsistent with a putative species tree, even after allowing discordance according to MSC. We model gene trees not as a set, but rather as a series (a realization of a stochastic process) along genomic positions. We propose a Hidden Markov Model (HMM) approach applied to quartet statistics measured from gene trees and tie the model to MSC using simulations. The combined use of these three ideas leads to a scalable method called Phlag. On simulated and real data, we show that Phlag can detect many cases of change in underlying evolutionary processes, including reduced recombination rates, population size changes, and admixture, all using the same algorithm. Phlag is available at github.com/bo1929/phlag. All results and scripts can be found at github.com/bo1929/shared.phlag.