Search Results (62)

Accurate estimation of the mean-reversion speed $\alpha$ in the AR(1) process $X_{t+1}=(1-\alpha )X_t+{\epsilon }_t$ is central to energy-commodity modelling. Classical estimators such as GARCH, jump-diffusion, and regime-switching produce model-conditioned estimates by embedding $\alpha$ within distributional assumptions, so that different model choices yield different $\widehat{\alpha }$ values from the same series without a principled criterion to adjudicate. We propose a distribution-free neural estimator based on a Temporal Convolutional Network (TCN) trained on synthetic AR(1) series with Sinh-ArcSinh (SAS) innovations. Distribution-free here means that no parametric family is assumed for the innovation distribution at inference time: the estimator imposes no distributional hypothesis when processing a new series. The SAS family serves as a training vehicle—not a model for the real data—chosen for its ability to span a broad range of tail weights and asymmetry profiles. The theoretical foundation is spectral invariance: the Yule–Walker equations establish that the autocorrelation structure ${\rho }_k={(1-\alpha )}^k$ depends on $\alpha$ alone, provided innovations are uncorrelated across lags—a condition satisfied not only by i.i.d. innovations but also by conditionally heteroscedastic processes such as GARCH. The TCN therefore generalises to volatility-clustering environments without modification, learning to extract $\alpha$ from temporal dependence alone, independently of the marginal innovation distribution and of the temporal variance structure. On held-out test series the estimator outperforms all classical competitors, with the advantage growing monotonically with non-Gaussianity. A robustness analysis on three out-of-distribution innovation families and on AR(1)-GARCH(1,1) processes empirically validates the spectral invariance guarantee across both marginal and temporal variance structure, including near-integrated GARCH processes where innovation kurtosis far exceeds the training range. The distribution-free $\widehat{\alpha }$ enables a two-stage pipeline in which $\alpha$ and the innovation distribution are characterised independently—a decoupling structurally impossible in classical likelihood-based approaches. Once trained, the TCN acts as a universal mean-reversion estimator applicable to any price series without re-fitting. Applied to four energy markets—Italian natural gas (PSV price), Italian electricity (PUN price), US Henry Hub, and US PJM West Hub—spanning log-return kurtosis from near-Gaussian to strongly heavy-tailed, the TCN yields robust, distribution-free estimates of mean-reversion speed. Full article

Prostate cancer is a major cause of cancer-related deaths among men in Sub-Saharan Africa, where late-stage diagnoses are common due to limited access to affordable and sensitive diagnostic tools. Early detection is essential to improve survival and reduce the disease burden. This review explores the integration of epigenetic biomarkers and CRISPR-Cas12a technology as a transformative approach for early, non-invasive prostate cancer detection in resource-limited settings. Among the many complexities of cancer development, molecular dysregulation plays a critical role. Epigenetic modifications including DNA methylation, histone changes, and non-coding RNA expression have emerged as stable and specific biomarkers with significant potential for the early detection and characterisation of prostate carcinogenesis. However, their low concentration in body fluids poses a significant challenge for detection. CRISPR-Cas12a, renowned for its high specificity and sensitivity, offers a promising solution. When integrated with isothermal amplification and liquid biopsy techniques, it enables rapid, point-of-care diagnostics. This review proposes a CRISPR-Cas12a-based diagnostic pipeline for the detection of specific epigenetic markers in liquid biopsies that could be associated with prostate cancer. The adoption of this technology in Sub-Saharan Africa has the potential to significantly improve early diagnosis, reduce mortality, and promote health equity. Full article

Remote photoplethysmography (rPPG) enables non-contact vital sign measurements using standard smart device cameras, opening up the potential of scalable health applications on consumer smart devices. However, rPPG signal quality is highly sensitive to camera sensor characteristics and image processing pipelines, which can vary between devices. This variation limits reproducibility and generalisation of rPPG-based algorithms beyond specific hardware platforms. This work presents a reproducible test environment for the validation of the camera sensor response in the context of rPPG signals. A microcontroller-driven illumination system and mechanically constrained setup are used to generate controlled, repeatable optical signals. Two characterisation tests are introduced: a time domain morphology analysis and a frequency domain attenuation analysis. Pulse timing consistency, pulse waveform morphology and normalised frequency responses are compared to assess sensor similarity. This method is applied to selected consumer devices and demonstrates consistent camera response patterns under the controlled test conditions. By explicitly addressing validation of the camera sensor and image processing pipeline, this work supports the development of more robust and transferable rPPG-based vital sign applications across a wider range of consumer devices. Full article

Background/Objectives: Milk somatic cells reflect the cellular composition and functional state of the lactating mammary gland and represent a valuable, non-invasive source for transcriptomic studies. Single-cell RNA sequencing (scRNA-seq) enables cell-type-resolved analysis of bovine milk; however, sequencing depth strongly influences the detection of lowly expressed genes and the resolution of transcriptional cell states. The aim of this study was to further characterise the single-cell transcriptome of bovine milk somatic cells, with particular emphasis on high-resolution gene expression profiling and cellular heterogeneity. Methods: Milk somatic cells were isolated from two healthy Holstein Friesian cows in mid-lactation and profiled using a droplet-based scRNA-seq platform. Newly generated high-depth datasets were integrated with two previously published bovine milk scRNA-seq datasets using an identical bioinformatics pipeline. Data integration, clustering and cell-type annotation were performed using the Seurat framework, and transcription factor expression was evaluated across datasets with different sequencing depths. Results: Single-cell transcriptomic analysis revealed a diverse cellular landscape in bovine milk, comprising epithelial, progenitor, and immune cell populations. Unsupervised clustering identified 21 transcriptionally distinct clusters, including multiple CD8⁺ T-cell subpopulations, monocytes, neutrophils, mast cells, and B cells, as well as luminal epithelial and luminal progenitor cells. While overall cell-type composition was comparable across datasets, deeply sequenced samples exhibited higher transcriptomic complexity and enabled refined resolution of immune and epithelial subpopulations. Deeper sequencing facilitated the detection of low-abundance transcription factors that were not observed in lower-depth datasets. Among these, NANOG was detected exclusively in deeply sequenced samples, suggesting the presence of rare transcriptional states associated with cellular plasticity. Conclusions: This study expands the single-cell transcriptomic landscape of bovine milk somatic cells and demonstrates the importance of sequencing depth for resolving functional cellular heterogeneity. The results highlight milk as a powerful, non-invasive source for investigating mammary gland biology and cellular plasticity during lactation. Full article

The aim of this study is to investigate and demonstrate the role of correction functions and optimisation-based refinement algorithms in enhancing the performance of machine learning models, particularly in predictive anomaly detection tasks applied in industrial environments. The performance of machine learning models is highly dependent on the quality of data preprocessing, model architecture, and post-processing methodology. In many practical applications—particularly in time-series forecasting and anomaly detection—the conventional training pipeline alone is insufficient, because model uncertainty, structural bias and the handling of rare events require specialised post hoc calibration and refinement mechanisms. This study provides a systematic overview of the role of correction functions (e.g., Principal Component Analysis (PCA), Squared Prediction Error (SPE)/Q-statistics, Hotelling’s T², Bayesian calibration) and adaptive improvement algorithms (e.g., Genetic Algorithms (GA), Particle Swarm Optimisation (PSO), Simulated Annealing (SA), Gaussian Mixture Model (GMM) and ensemble-based techniques) in enhancing the performance of machine learning pipelines. The models were trained on a real industrial dataset compiled from power network analytics and harmonic-injection-based loading conditions. Model validation and equipment-level testing were performed using a large-scale harmonic measurement dataset collected over a five-year period. The reliability of the approach was confirmed by comparing predicted state transitions with actual fault occurrences, demonstrating its practical applicability and suitability for integration into predictive maintenance frameworks. The analysis demonstrates that correction functions introduce deterministic transformations in the data or error space, whereas improvement algorithms apply adaptive optimisation to fine-tune model parameters or decision boundaries. The combined use of these approaches significantly reduces overfitting, improves predictive accuracy and lowers false alarm rates. This work introduces the concept of an Organically Adaptive Predictive (OAP) ML model. The proposed model presents organic adaptivity, continuously adjusting its predictive behaviour in response to dynamic variations in network loading and harmonic spectrum composition. The introduced terminology characterises the organically emergent nature of the adaptive learning mechanism. Full article

Retrieval-Augmented Generation (RAG) has been observed to encounter challenges in heterogeneous query scenarios characterised by varying evidence requirements and reasoning depths. In order to address this limitation, the present paper puts forward a proposal for an Adaptive Multi-Source RAG framework (AMSRAG) that integrates query complexity awareness with confidence-aware fusion. The framework performs query complexity classification with a pretrained language model, calibrates the classification confidence to guide the dynamic scheduling of retrieval paths and the adjustment of fusion weights, and enables a controllable balance between answer quality and retrieval efficiency through hierarchical path selection and cross-source weighting. The experiments conducted on multiple open-domain question-answering datasets demonstrate that the query complexity classifier achieves an accuracy of 85.9% and a Macro-F1 score of 85.4%. These outcomes indicate the potential for the classifier to generate a reliable decision signal, which can subsequently be utilised to guide the process of adaptive retrieval and fusion. The proposed framework demonstrates a marked improvement in terms of both answer accuracy and retrieval relevance when compared to the fixed-pipeline RAG. In scenarios involving high-confidence queries, the system has been shown to effectively avoid redundant retrieval, thereby reducing the average number of retrievals. In instances of low-confidence complex queries, the system has been shown to enhance evidence coverage and completeness of answers through multi-source retrieval and confidence-weighted fusion. This study proposes a novel methodology for enhancing the adaptability and resource efficiency of RAG systems in response to heterogeneous query conditions. Full article

Background: Three-finger toxins (3FTxs) are a major axis of functional diversification in advanced snake venoms, with canonical paralytic activity mediated through muscle-type nicotinic acetylcholine receptors (nAChRs) and a broader set of non-nicotinic targets. This review integrates evidence bearing on coevolution between 3FTxs and target receptors, spanning toxin origin, diversification, receptor evolution, and ecological context. Methods: The synthesis draws on comparative genomic and transcriptomic studies of 3FTx gene-family evolution, codon-model analyses of selection, structural characterisation of toxin–receptor interfaces, and functional assays (including receptor-mimicking peptide binding) that link sequence variation to binding and toxicity. Results: Across lineages, 3FTx diversification is repeatedly structured by strong constraint on the disulphide-rich scaffold with accelerated change concentrated in solvent-exposed loops, alongside birth–death dynamics and exon/segment-level innovation that expand binding specificity. On the receptor side, resistance-associated variation is most intensively characterised for the nAChR α₁ orthosteric site and includes convergent, mechanistically distinct solutions such as electrostatic repulsion and glycosylation-mediated steric interference. Within the predominantly elapid systems currently examined, integrative datasets indicate that prey-selective binding and geographically variable susceptibility can arise from modest substitutions at toxin–receptor interfaces, but they also reveal substantial taxonomic and target-specific biases. Conclusions: Current evidence supports adaptive diversification in both toxins and receptors, while broader evolutionary interpretations are limited by uneven sampling and the frequent lack of matched toxin and receptor variants analysed within a common evolutionary framework. Development of predictive models will require joint pipelines linking genomics, structure-informed evolutionary inference, scalable functional assays, and explicit ecological network context. Full article

Background/Objectives: Preclinical models that accurately reflect Ewing sarcoma (ES) will enable the prioritisation of clinically active targeted agents from bench to clinic. To expedite this process, we have established and characterised patient-derived ES cultures (PDES) in vitro. Methods: Fluorescence in situ hybridisation, RT-PCR and western blotting were used to examine expression of the pathognomonic EWSR1 fusions. Activation or repression of EWSR1 fusion downstream targets and proliferation was examined by immunofluorescence and immunohistochemistry. Using next-generation sequencing, the DNA and transcriptomic profiles of PDES and cell lines were compared. The response of PDES and cell lines to standard-of-care chemotherapeutics, ionising radiation and investigational drugs was examined. Results: All PDES contain EWSR1 fusion DNA, consistent with a diagnosis of ES. EWSR1 fusion gene RNA and protein were detected in 70% and 21% of PDES, respectively. Markers of proliferation and expression of EWSR1 fusion target genes were consistent with the tumours from which PDES were derived (R² = 0.74, p < 0.0001) and the paediatric mesenchymal lineage (SBS5 and SBS1, ID1 and ID2). In contrast, the transcriptome of PDES was significantly different from that of cell lines. PDES had a significantly increased doubling time (p < 0.00001), decreased expression of Ki67 (p < 0.0001) and increased migration (p < 0.02) compared to cell lines. Consistent with the longer doubling time, PDES were more resistant to doxorubicin, etoposide and vincristine and ionising radiation (p < 0.0001) than cell lines. PDES were sensitive to mTKIs (cabozantinib, lenvatinib, and regorafenib), and trabectedin. The response of PDES to drugs in vitro reflects the clinical experience of patients. Conclusions: Models incorporating PDES cells may positively contribute to the preclinical pipeline. Full article

Persistence of seagrass meadows varies depending on community composition, substrate stability, environmental forcing, and water quality/clarity. Spatial trends in decadal scale persistence are difficult to assess at the meadow scale using in situ approaches and assessments using Earth Observation often lack temporal consistency. This study utilises a multi-decadal field monitoring dataset and high-resolution multispectral satellite imagery in a cloud-processing environment to assess species distribution, seagrass cover, and meadow persistence. In this work, we investigate long-term trends in overall meadow and species-specific persistence in the Eastern Banks, Moreton Bay, Australia, a shallow, semi-enclosed, subtropical embayment (∼200 km²). Here, we have identified an overall decline in seagrass cover (−15% of the total study area), between 2011 and 2025, through contraction of meadow extent, with most losses in colonising species (Halophila spinulosa and Halophila ovalis) across the deeper sections of the study area. We have also quantified the spatial extent of a previously identified broad-scale ecosystem shift from meadows dominated by Zostera muelleri to a prevalence of Oceana serrulata, and reduction in the sparse cover species H. spinulosa and H. ovalis. We have presented a semi-automated cloud-processing based pipeline to combine in situ seagrass observations, derived from an expertly trained machine learning model, with high resolution multispectral data to assess seagrass cover and persistence. The variability in decadal-scale persistence between the six key species found in this region has been assessed, with dense cover species (e.g., O. serrulata and Z. muelleri) exhibiting moderate persistence (>0.32) and sparse cover species (H. ovalis and H. spinulosa) with low persistence (∼0.15). Colonising/opportunistic growth patterns characterise the species examined in this study, indicating quick response to disturbance but a lack temporal consistency in meadow form, which has critical implications for resilience. Full article

Dents are among the most common deformation defects in buried transmission pipelines, significantly influencing structural integrity and internal flow behaviour. This study examines the occurrence of turbulence in dented pipe sections using time–frequency analysis of acoustic emission (AE) responses. The approach aims to overcome the challenge of obtaining meaningful information from AE signals during conventional dent inspections. By correlating AE spectral characteristics with flow-induced turbulence, the study provides insights into how mechanical deformation influences AE signal behaviour, contributing to an improved assessment of pipeline integrity. In this study, AE signals were captured during flow loop tests on healthy, 5%, 15%, and 30% dented pipe sections to evaluate the influence of dent severity on turbulence behaviour. Time–frequency domain analysis using the Morlet wavelet transform on the starting, middle, and end segments of AE signals revealed a progressive increase in signal energy with increasing dent depth, reaching a maximum of 2.54 × 10⁻⁰⁸ μE²/Hz − 2.54 × 10⁻⁰⁸ μE²/Hz for the end segment of AE signals under the 30% dented pipe condition. Complementary computational fluid dynamics (CFD) simulations were performed to compute velocity streamlines and corresponding Reynolds numbers for validating the turbulence detection results. A strong correlation between the CWT coefficient energy and Reynolds number, with R² values of 0.9633, 0.9007, and 0.9052 for the starting, middle, and end signal segments, respectively, was observed. These findings demonstrate that AE time–frequency analysis offers a reliable diagnostic approach for identifying and characterising dent-induced turbulence in pipeline systems. Full article

Partial blockages in pressurised pipe systems present significant challenges for precise detection, characterisation, and ongoing monitoring. Transient test-based techniques, which utilise sharp but small pressure waves, have shown considerable potential due to their safety and diagnostic capabilities. This paper investigates the transient response of an extended partial blockage—an evolution of a discrete partial blockage that protrudes longitudinally—an increasingly complex condition which has a greater impact on the behavior of pipe systems. Through Computational Fluid Dynamics simulations, the interaction of pressure waves with extended partial blockages of different severity and lengths is examined to assess the resulting pressure response. The results confirm that the pressure signature, generated by extended partial blockages, differs markedly from those of discrete partial blockages. In particular, the magnitudes of the first and second pressure peaks enable accurate characterisation of the severity and extent of the extended partial blockage. These results demonstrate that transient test-based techniques can play a significant role in managing water pipe systems, facilitating more targeted maintenance interventions. Broader implementation of these techniques could enable water utilities to reduce energy consumption, maintain water quality with lower chlorine dosing, and prevent the progression of partial blockages to total pipeline blockage. Full article

Genome sequencing has possibly been the greatest step in the development of advanced tools for animal genetic improvement: knowledge of gene sequences and use of haplotype markers for productivity traits can provide important improvements in yield production and optimisation of reproductive program. Next-generation and, more recently, third-generation sequencing techniques enormously increased the ability to produce sequences from single individuals and increased the interest in exome or whole-genome sequencing as an alternative to SNP chips in breeding programs as these techniques allowed for the capture of a wider range of variations, including characterisation of rare variants, structural variations, and copy number changes. Here, we present a procedure, based on fast de novo assembly and a scaffolding step, to quickly build an almost complete genome starting from long reads obtained in a single sequencing run. The procedure, applied to sequences from five water buffaloes, was able to independently build, for each individual, an almost complete high-quality genome with highly continuous chromosome sequences; in most cases, over 90% of the length of the reference chromosome was covered by less than ten long contigs. Unlike other pipelines based on slower assemblers or which require many sequencing data, in 1–2 days, the proposed procedure can go from a single run to continuous genome assembly, supporting fast analysis of large chromosome structures, potentially useful for improving animal breeding and productivity. Full article

Viral metagenomic next-generation sequencing (vmNGS) has transformed our capacity for the untargeted detection and characterisation of (re)emerging zoonotic viruses, surpassing the limitations of traditional targeted diagnostics. In this review, we critically evaluate the current landscape of vmNGS, highlighting its integration within the One Health paradigm and its application to the surveillance and discovery of (re)emerging viruses at the human–animal–environment interface. We provide a detailed overview of vmNGS workflows including sample selection, nucleic acid extraction, host depletion, virus enrichment, sequencing platforms, and bioinformatic pipelines, all tailored to maximise sensitivity and specificity for diverse sample types. Through selected case studies, including SARS-CoV-2, mpox, Zika virus, and a novel henipavirus, we illustrate the impact of vmNGS in outbreak detection, genomic surveillance, molecular epidemiology, and the development of diagnostics and vaccines. The review further examines the relative strengths and limitations of vmNGS in both passive and active surveillance, addressing barriers such as cost, infrastructure requirements, and the need for interdisciplinary collaboration. By integrating molecular, ecological, and public health perspectives, vmNGS stands as a central tool for early warning, comprehensive monitoring, and informed intervention against (re)emerging viral threats, underscoring its critical role in global pandemic preparedness and zoonotic disease control. Full article

Background: HM is a rare, severe form of migraine with aura, characterised by motor weakness and strongly influenced by genetic factors affecting the brain. While pathogenic variants in CACNA1A, ATP1A2, and SCN1A genes have been implicated in familial HM, approximately 75% of cases lack known pathogenic variants in these genes, suggesting a more complex genetic basis. Methods: To advance our understanding of HM, we applied a variant prioritisation approach using whole-exome sequencing (WES) data from patients referred for HM diagnosis (n = 184) and utilised PathVar, a bioinformatics pipeline designed to identify pathogenic variants. Our analysis incorporated two strategies for association testing: (1) PathVar-identified single nucleotide variants (SNVs) and (2) PathVar SNVs combined with missense and rare variants. Principal component analysis (PCA) was performed to adjust for ancestral and other unknown differences between cases and controls. Results: Our results reveal a sequential reduction in the number of genes significantly associated with HM, from 20 in the first strategy to 11 in the second, which highlights the unique contribution of PathVar SNVs to the genetic architecture of HM. PathVar SNVs were more distinctive in the case cohort, suggesting a closer link to the functional changes underlying HM compared to controls. Notably, novel genes, such as SLC38A10, GCOM1, and NXPH2, which were previously not implicated in HM, are now associated with the disorder, advancing our understanding of its genetic basis. Conclusions: By prioritising PathVar SNVs, we identified a broader set of genes potentially contributing to HM. Given that HM is a rare condition, our findings, utilising a sample size of 184, represent a unique contribution to the field. This iterative analysis demonstrates that integrating diverse variant schemes provides a more comprehensive view of the genetic factors driving HM. Full article

Third-generation sequencing (TGS), also known as long-read sequencing, has become a promising tool in clinical and research laboratories because it delivers high-resolution results with unmatched throughput. Specialised immunohematology laboratories currently employ sequencing-based methods to characterise rare ABO blood group phenotypes that cannot be identified through serology and genotyping methods. However, routine clinical application of these methods remains elusive due to the absence of validated laboratory protocols and bioinformatics tools. In this study, we have developed and validated a TGS-based workflow for comprehensive determination of the clinically relevant ABO phenotypes from DNA isolated from buccal swabs or whole blood. The region spanning exons 2 to 7 of the ABO gene were amplified and sequenced on MinION 10.4.1 flow cells. Predicted ABO phenotypes were initially determined based on single-nucleotide variants at gDNA261 (rs8176719), gDNA796 (rs8176746), and gDNA803 (rs8176747). However, certain O subtypes lacked the distinguishing deletion (rs8176719) and instead exhibited variations in exon 7 at gDNA802 (rs41302905) and gDNA805, caused by gDNA804 (rs782782485), which differentiate them from A alleles sharing the same nucleotides at gDNA261, gDNA796, and gDNA803. These additional variants were added to the analysis pipeline to identify the additional subtypes. DNA sequence data were sufficient to distinguish between the four clinically relevant ABO blood group phenotypes based on five polymorphic positions. While high sequencing coverage allowed for higher resolution genetic analysis, as few as 20 reads are sufficient for determining the ABO genotype and predicted phenotype of an individual. Typing results generated by this pipeline showed remarkable concordance with both serological results and molecular typing results by an independent laboratory, indicating its accuracy and reliability. This study demonstrates a comprehensive characterisation of clinically relevant ABO blood genotypes and predicted phenotypes using TGS methods. The approach provided a scalable and precise method for routine ABO blood group screening and aided in the development of pioneering bioinformatics tools suitable for clinical and research application. Full article