Development of a deep-learning model tailored for HER2 detection in breast cancer to aid pathologists in interpreting HER2-low cases.

AIMS: Over 50% of breast cancer cases are "Human epidermal growth factor receptor 2 (HER2) low breast cancer (BC)", characterized by HER2 immunohistochemistry (IHC) scores of 1+ or 2+ alongside no amplification on fluorescence in situ hybridization (FISH) testing. The development of new anti-HER2 antibody-drug conjugates (ADCs) for treating HER2-low breast cancers illustrates the importance of accurately assessing HER2 status, particularly HER2-low breast cancer. In this study we evaluated the performance of a deep-learning (DL) model for the assessment of HER2, including an assessment of the causes of discordances of HER2-Null between a pathologist and the DL model. We specifically focussed on aligning the DL model rules with the ASCO/CAP guidelines, including stained cells' staining intensity and completeness of membrane staining. METHODS AND RESULTS: We trained a DL model on a multicentric cohort of breast cancer cases with HER2-IHC scores (n = 299). The model was validated on two independent multicentric validation cohorts (n = 369 and n = 92), with all cases reviewed by three senior breast pathologists. All cases underwent a thorough review by three senior breast pathologists, with the ground truth determined by a majority consensus on the final HER2 score among the pathologists. In total, 760 breast cancer cases were utilized throughout the training and validation phases of the study. The model's concordance with the ground truth (ICC = 0.77 [0.68-0.83]; Fisher P = 1.32e-10) is higher than the average agreement among the three senior pathologists (ICC = 0.45 [0.17-0.65]; Fisher P = 2e-3). In the two validation cohorts, the DL model identifies 95% [93% - 98%] and 97% [91% - 100%] of HER2-low and HER2-positive tumours, respectively. Discordant results were characterized by morphological features such as extended fibrosis, a high number of tumour-infiltrating lymphocytes, and necrosis, whilst some artefacts such as nonspecific background cytoplasmic stain in the cytoplasm of tumour cells also cause discrepancy. CONCLUSION: Deep learning can support pathologists' interpretation of difficult HER2-low cases. Morphological variables and some specific artefacts can cause discrepant HER2-scores between the pathologist and the DL model.

Evolutionary Strategies Enable Systematic and Reliable Uncertainty Quantification: A Proof-of-Concept Pilot Study on Resting-State Functional MRI Language Lateralization.

Reliable and trustworthy artificial intelligence (AI), particularly in high-stake medical diagnoses, necessitates effective uncertainty quantification (UQ). Existing UQ methods using model ensembles often introduce invalid variability or computational complexity, rendering them impractical and ineffective in clinical workflow. We propose a UQ approach based on deep neuroevolution (DNE), a data-efficient optimization strategy. Our goal is to replicate trends observed in expert-based UQ. We focused on language lateralization maps from resting-state functional MRI (rs-fMRI). Fifty rs-fMRI maps were divided into training/testing (30:20) sets, representing two labels: "left-dominant" and "co-dominant." DNE facilitated acquiring an ensemble of 100 models with high training and testing set accuracy. Model uncertainty was derived from distribution entropies over the 100 model predictions. Expert reviewers provided user-based uncertainties for comparison. Model (epistemic) and user-based (aleatoric) uncertainties were consistent in the independently and identically distributed (IID) testing set, mainly indicating low uncertainty. In a mostly out-of-distribution (OOD) holdout set, both model and user-based entropies correlated but displayed a bimodal distribution, with one peak representing low and another high uncertainty. We also found a statistically significant positive correlation between epistemic and aleatoric uncertainties. DNE-based UQ effectively mirrored user-based uncertainties, particularly highlighting increased uncertainty in OOD images. We conclude that DNE-based UQ correlates with expert assessments, making it reliable for our use case and potentially for other radiology applications.

Quantitative analysis of solid dosage forms of Atenolol by Raman spectroscopy.

ObjectiveTo develop a Raman spectroscopy-based analytical model for quantification of solid dosage forms of active pharmaceutical ingredient (API) of Atenolol.Significance:For the quantitative analysis of pharmaceutical drugs, Raman Spectroscopy is a reliable and fast detection method. As part of this study, Raman Spectroscopy is explored for the quantitative analysis of different concentrations of Atenolol.MethodsVarious solid-dosage forms of Atenolol were prepared by mixing API with excipients to form different solid-dosage formulations of Atenolol. Multivariate data analysis techniques such as Principal Component Analysis (PCA) and Partial least square regression (PLSR) were used for the qualitative and quantitative analysis, respectively.ResultsAs the concentration of the drug increased in formulation, the peak intensities of the distinctive Raman spectral characteristics associated with the API (Atenolol) gradually increased. Raman spectral data sets were classified using PCA due to their distinctive spectral characteristics. Additionally, a prediction model was built using PLSR analysis to assess the quantitative relationship between various API (Atenolol) concentrations and spectral features. With a goodness of fit value of 0.99, the root mean square errors of calibration (RMSEC) and prediction (RMSEP) were determined to be 1.0036 mg and 2.83 mg, respectively. The API content in the blind/unknown Atenolol formulation was determined as well using the PLSR model.ConclusionBased on these results, Raman spectroscopy may be used to quickly and accurately analyze pharmaceutical samples and for their quantitative determination.

Reducing overconfident errors in molecular property classification using Posterior Network.

Deep-learning-based classification models are increasingly used for predicting molecular properties in drug development. However, traditional classification models using the Softmax function often give overconfident mispredictions for out-of-distribution samples, highlighting a critical lack of accurate uncertainty estimation. Such limitations can result in substantial costs and should be avoided during drug development. Inspired by advances in evidential deep learning and Posterior Network, we replaced the Softmax function with a normalizing flow to enhance the uncertainty estimation ability of the model in molecular property classification. The proposed strategy was evaluated across diverse scenarios, including simulated experiments based on a synthetic dataset, ADMET predictions, and ligand-based virtual screening. The results demonstrate that compared with the vanilla model, the proposed strategy effectively alleviates the problem of giving overconfident but incorrect predictions. Our findings support the promising application of evidential deep learning in drug development and offer a valuable framework for further research.

Quantitative PCR-based high-sensitivity detection of HBV-DNA levels reflects liver function deterioration in patients with hepatitis B virus-related cirrhosis.

OBJECTIVES: To investigate the clinical implication of quantitative polymerase chain reaction (PCR)-based high-sensitivity detection of hepatitis B virus (HBV)-DNA levels in patients with HBV-related liver cirrhosis (LC). METHODS: From January 2020 to December 2022, 100 fasting serum samples were collected and retrospectively analyzed from patients with treated HBV-related LC attending the Suzhou Hospital of Integrated Traditional Chinese and Western Medicine and Suzhou Guangci Cancer Hospital. Patients were divided into a negative group (HBV-DNA < 20 IU/mL) and a positive group (HBV-DNA ≥ 20 IU/mL) according to their high-sensitivity HBV-DNA test results. The clinical characteristics and serological indicators of the two groups were compared, mainly including gender, age, liver function [total protein (TP), albumin (ALB), alanine aminotransferase (ALT), aspartate aminotransferase (AST), γ-glutamyl transpeptidase (GGT), alkaline phosphatase (ALP), total bilirubin (TBIL), direct bilirubin (DBIL), and indirect bilirubin (IBIL)], lipids [total cholesterol (TC) and triglycerides (TG)], platelets (PLT), five serum liver fibrosis markers [cholyglycine (CG), hyaluronic acid (HA), laminin (LN), precollagen type III (PCIII), and type IV collagen (IV-C)], serum gastrointestinal tumor markers [α-fetoprotein (AFP) and carcinoembryonic antigen (CEA)], and hepatitis B surface antigen (HBsAg). The differences between the two groups in terms of liver function Child-Pugh grades and the incidence of hepatocellular carcinoma (HCC) were also compared. RESULTS: There were 39 patients in the positive group, including 29 males and 10 females, and 61 patients in the negative group, including 38 males and 23 females, with no statistically significant differences in gender and age distribution between the two groups (P > 0.05). The levels of serological indicators (TP, ALB, AST, GGT, ALP, TBIL, DBIL, IBIL, TC, TG, PLT, CG, HA, LN, PCIII, IV-C, AFP, CEA, and HBsAg) in both groups showed no significant differences (P > 0.05), but the ALT level in the positive group was higher than that in the negative group (P < 0.0001). The positive group had worse Child-Pugh grades and higher HCC incidence compared to the negative group (P < 0.0001, P = 0.028). CONCLUSIONS: Patients with HBV-related LC and HBV-DNA ≥ 20 IU/mL have higher serum ALT levels, worse liver function Child-Pugh grades, and higher HCC incidence than those with HBV-DNA < 20 IU/mL. High-sensitivity HBV-DNA quantification can reflect the deterioration of liver function in patients with HBV-related LC to some extent.

Fast neonicotinoid quantification in honey using the one-point internal calibration approach.

Neonicotinoids, a highly effective class of insecticides used worldwide, have been identified as a major cause of concern for biodiversity. To assess the ecological and environmental consequences of neonicotinoids' use, reliable analytical methodologies, including calibration approaches, are needed. Here, we compared the performance of internal calibration (IC) using a single concentration of stable isotope-labeled standard (SIL) with classical multipoint external calibration (EC) for the quantification of six neonicotinoids in honey. IC showed acceptable levels of trueness (86.3% - 116.0%) and precision (1.4% - 20.8%), although slight biases were observed at very low concentrations compared to EC. When applied to 32 original honey samples, both approaches showed strong agreement (R2 > 0.998) with proportional biases lower than 5%. These results highlight the possibility of implementing IC to simplify quantification in liquid chromatography-mass spectrometry-based pesticide applications.

Improving brain atrophy quantification with deep learning from automated labels using tissue similarity priors.

Brain atrophy measurements derived from magnetic resonance imaging (MRI) are a promising marker for the diagnosis and prognosis of neurodegenerative pathologies such as Alzheimer's disease or multiple sclerosis. However, its use in individualized assessments is currently discouraged due to a series of technical and biological issues. In this work, we present a deep learning pipeline for segmentation-based brain atrophy quantification that improves upon the automated labels of the reference method from which it learns. This goal is achieved through tissue similarity regularization that exploits the a priori knowledge that scans from the same subject made within a short interval must have similar tissue volumes. To train the presented pipeline, we use unlabeled pairs of T1-weighted MRI scans having a tissue similarity prior, and generate the target brain tissue segmentations in a fully automated manner using the fsl_anat pipeline implemented in the FMRIB Software Library (FSL). Tissue similarity regularization is enforced during training through a weighted loss term that penalizes tissue volume differences between short-interval scan pairs from the same subject. In inference, the pipeline performs end-to-end skull stripping and brain tissue segmentation from a single T1-weighted MRI scan in its native space, i.e., without performing image interpolation. For longitudinal evaluation, each image is independently segmented first, and then measures of change are computed. We evaluate the presented pipeline in two different MRI datasets, MIRIAD and ADNI1, which have longitudinal and short-interval imaging from healthy controls (HC) and Alzheimer's disease (AD) subjects. In short-interval scan pairs, tissue similarity regularization reduces the quantification error and improves the consistency of measured tissue volumes. In the longitudinal case, the proposed pipeline shows reduced variability of atrophy measures and higher effect sizes of differences in annualized rates between HC and AD subjects. Our pipeline obtains a Cohen's d effect size of d=2.07 on the MIRIAD dataset, an increase from the reference pipeline used to train it (d=1.01), and higher than that of SIENA (d=1.73), a well-known state-of-the-art approach. In the ADNI1 dataset, the proposed pipeline improves its effect size (d=1.37) with respect to the reference pipeline (d=0.80) and surpasses SIENA (d=1.33). The proposed data-driven deep learning regularization reduces the biases and systematic errors learned from the reference segmentation method, which is used to generate the training targets. Improving the accuracy and reliability of atrophy quantification methods is essential to unlock brain atrophy as a diagnostic and prognostic marker in neurodegenerative pathologies.

Risk management and loss prevention strategies for fertilizer industries.

Various toxic and flammable gases exist in the fertilizer industry whose release quantification is very important regarding emergency preparedness, planning and response, and well-being of the community. ALOHA threat zones and threat at a point coupled with MARPLOT are evaluated for ammonia, methane, carbon dioxide and hydrogen release, and outdoor and indoor concentrations of these gases in nearby residences and highways calculated. These footprints are calculated using ALOHA which requires inputs such as site data, site location, building type, gas name, atmospheric inputs, release source information and dispersion model to display the threat zone, which can then be shown on MARPLOT. Potential impact of these releases on the community is mitigated through releasing equipment isolations, water sprays for dilutions, dilutions through steam or air and emergency sirens for information. This article covers hazards in the fertilizer industry, and provides general guidelines for operational staff of any industry to mitigate hazards.

NMR analysis, cytotoxic activity and theoretical study of a complex between SRPIN340 and p-sulfonic acid calix[6]arene.

Aim: This study aimed to enhance the aqueous dissolution of SRPK inhibitor N-(2-(piperidin-1-yl)-5-(trifluoromethyl)phenyl)isonicotinamide (SRPIN340). Materials & Methods: A complex with p-sulfonic calix[6]arene (Host) and SRPIN340 (Guest) was prepared, studied via 1H nuclear magnetic resonance (NMR) and theoretical calculations and biologically evaluated on cancer cell lines. Results & conclusion: The 1:1 host (H)/guest (G) complex significantly enhanced the aqueous dissolution of SRPIN340, achieving 64.8% water solubility as determined by 1H NMR quantification analysis. The H/G complex reduced cell viability by 75% for HL60, ∼50% for Nalm6 and Jurkat, and ∼30% for B16F10 cells. It exhibited greater cytotoxicity than free SRPIN340 against Jurkat and B16F10 cells. Theoretical studies indicated hydrogen bond stabilization of the complex, suggesting broader applicability of SRPIN340 across diverse biological systems.

[Box: see text].

Super-resolution Multispectral Imaging Nanoimmunosensor for Simultaneous Detection of Diverse Early Cancer Biomarkers.

In medical diagnosis, relying on only one type of biomarker is insufficient to accurately identify cancer. Blood-based multicancer early detection can help identify more than one type of cancer from a single blood sample. In this study, a super-resolution multispectral imaging nanoimmunosensor (srMINI) based on three quantum dots (QDs) of different color conjugated with streptavidin was developed for the simultaneous screening of various cancer biomarkers in blood at the single-molecule level. In the experiment, the srMINI chip was used to simultaneously detect three key cancer biomarkers: carcinoembryonic antigen (CEA), C-reactive protein (CRP), and alpha-fetoprotein (AFP). The srMINI chip exhibited 108 times higher detection sensitivity of 0.18-0.5 ag/mL (1.1-2.6 zM) for these cancer biomarkers than commercial enzyme-linked immunosorbent assay kits because of the absence of interfering signals from the substrate, establishing considerable potential for multiplex detection of cancer biomarkers in blood. Therefore, the simultaneous detection of various cancer biomarkers using the developed srMINI chip with high diagnostic precision and accuracy is expected to play a decisive role in early diagnosis or community screening as a single-molecule biosensor.

Identifying Bayesian optimal experiments for uncertain biochemical pathway models.

Pharmacodynamic (PD) models are mathematical models of cellular reaction networks that include drug mechanisms of action. These models are useful for studying predictive therapeutic outcomes of novel drug therapies in silico. However, PD models are known to possess significant uncertainty with respect to constituent parameter data, leading to uncertainty in the model predictions. Furthermore, experimental data to calibrate these models is often limited or unavailable for novel pathways. In this study, we present a Bayesian optimal experimental design approach for improving PD model prediction accuracy. We then apply our method using simulated experimental data to account for uncertainty in hypothetical laboratory measurements. This leads to a probabilistic prediction of drug performance and a quantitative measure of which prospective laboratory experiment will optimally reduce prediction uncertainty in the PD model. The methods proposed here provide a way forward for uncertainty quantification and guided experimental design for models of novel biological pathways.

Quantifying functional redundancy in polysaccharide-degrading prokaryotic communities.

BACKGROUND: Functional redundancy (FR) is widely present, but there is no consensus on its formation process and influencing factors. Taxonomically distinct microorganisms possessing genes for the same function in a community lead to within-community FR, and distinct assemblies of microorganisms in different communities playing the same functional roles are termed between-community FR. We proposed two formulas to respectively quantify the degree of functional redundancy within and between communities and analyzed the FR degrees of carbohydrate degradation functions in global environment samples using the genetic information of glycoside hydrolases (GHs) encoded by prokaryotes. RESULTS: Our results revealed that GHs are each encoded by multiple taxonomically distinct prokaryotes within a community, and the enzyme-encoding prokaryotes are further distinct between almost any community pairs. The within- and between-FR degrees are primarily affected by the alpha and beta community diversities, respectively, and are also affected by environmental factors (e.g., pH, temperature, and salinity). The FR degree of the prokaryotic community is determined by deterministic factors. CONCLUSIONS: We conclude that the functional redundancy of GHs is a stabilized community characteristic. This study helps to determine the FR formation process and influencing factors and provides new insights into the relationships between prokaryotic community biodiversity and ecosystem functions. Video Abstract.

ICP-MS As a Contributing Tool to Nontarget Screening (NTS) Analysis for Environmental Monitoring.

Due to the increasing number of chemicals released into the environment, nontarget screening (NTS) analysis is a necessary tool for providing comprehensive chemical analysis of environmental pollutants. However, NTS workflows encounter challenges in detecting both known and unknown pollutants with common chromatography high-resolution mass spectrometry (HRMS) methods. Identification of unknowns is hindered by limited elemental composition information, and quantification without identical reference standards is prone to errors. To address these issues, we propose the use of inductively coupled plasma mass spectrometry (ICP-MS) as an element-specific detector. ICP-MS can enhance the confidence of compound identification and improve quantification in NTS due to its element-specific response and unambiguous chemical composition information. Additionally, mass balance calculations for individual elements (F, Br, Cl, etc.) enable assessment of total recovery of those elements and evaluation of NTS workflows. Despite its benefits, implementing ICP-MS in NTS analysis and environmental regulation requires overcoming certain shortcomings and challenges, which are discussed herein.

Ultrasensitive detection of 5-hydroxymethylcytosine in genomic DNA using a graphene-based sensor modified with biotin and gold nanoparticles.

Ten-eleven translocation (TET) proteins orchestrate deoxyribonucleic acid (DNA) methylation-demethylation dynamics by oxidizing 5-methylcytosine to 5-hydroxymethylcytosine (5hmC) and are frequently inactivated in various cancers. Due to the significance of 5hmC as an epigenetic biomarker for cancer diagnosis, pathogenesis, and treatment, its rapid and precise quantification is essential. Here, we report a highly sensitive electrochemical method for quantifying genomic 5hmC using graphene sheets that were electrochemically exfoliated and functionalized with biotin and gold nanoparticles (Bt-AuNPs) through a single-step electrical method. The attachment of Bt-AuNPs to graphene enhances the specificity of 5hmC-containing DNA and augments the oxidation of 5hmC to 5-formylcytosine in DNA. When coupled to a gold electrode, the Bt-AuNP-graphene-based sensor exhibits exceptional sensitivity and specificity for detecting 5hmC, with a detection limit of 63.2 fM. Furthermore, our sensor exhibits a remarkable capacity to measure 5hmC levels across a range of biological samples, including preclinical mouse tissues with varying 5hmC levels due to either TET gene disruption or oncogenic transformation, as well as human prostate cancer cell lines. Therefore, our sensing strategy has substantial potential for cancer diagnostics and prognosis.

Quantifying the effects of five rehabilitation training methods on the ability of elderly men to control bowel movements: a finite element analysis study.

Purpose: The study aims to develop a finite element model of the pelvic floor and thighs of elderly men to quantitatively assess the impact of different pelvic floor muscle trainings and the urinary and defecation control ability. Methods: A finite element model of the pelvic floor and thighs of elderly men was constructed based on MRI and CT. Material properties of pelvic floor tissues were assigned through literature review, and the relative changes in waistline, retrovesical angle (RVA) and anorectad angulation (ARA) to quantitatively verify the effectiveness of the model. By changing the material properties of muscles, the study analyzed the muscle strengthening or impairment effects of the five types of rehabilitation training for four types of urination and defecation dysfunction. The changes in four outcome indicators, including the retrovesical angle, anorectad angulation, stress, and strain, were compared. Results: This study indicates that ARA and RVA approached their normal ranges as material properties changed, indicating an enhancement in the urinary and defecation control ability, particularly through targeted exercises for the levator ani muscle, external anal sphincter, and pelvic floor muscles. This study also emphasizes the effectiveness of personalized rehabilitation programs including biofeedback, exercise training, electrical stimulation, magnetic stimulation, and vibration training and advocates for providing optimized rehabilitation training methods for elderly patients. Discussion: Based on the results of computational biomechanics, this study provides foundational scientific insights and practical recommendations for rehabilitation training of the elderly's urinary and defecation control ability, thereby improving their quality of life. In addition, this study also provides new perspectives and potential applications of finite element analysis in elderly men, particularly in evaluating and designing targeted rehabilitation training.

Lumbar and Cervical Concurrent Muscle Activity Analysis During Hamstring Length Testing to Design a Customized Treatment Plan: An Observational Cross-Sectional Study.

BACKGROUND:  Hamstring length plays a significant role in a spectrum of clinical entities, from injury prevention and gait dysfunction to posture correction. Evidence suggests that the prevalence of hamstring tightness (HT)/reduced length is increasing. Despite the number of available tests and treatment protocols, HT is still a functional diagnosis. This study's primary goal is to evaluate concurrent muscle (CM) usage during these testing procedures to design a unique, customized treatment protocol. METHODS: The study was conducted in two stages. In phase 1, Active Straight Leg Raise (ASLR), Active Total Knee Extension (ATKE), and Active Seated Total Knee Extension (ASTKE) were carried out. Next, two pressure gauges (PGs) were placed to align with the natural lumbar and cervical curvatures while testing ASLR and ATKE. After analyzing the results for pressure gauge placement, phase 2 data were collected for tests ASLR and ATKE with PG. RESULTS: The results of ASLR and ATKE, both with and without PG, indicate a high prevalence rate, whereas the results of ASTKE show no prevalence. Changes in the PG values indicate CM usage. Dichotomization revealed that participants with normal test scores (non-HT group) had increased usage of CM work. Positive and negative changes in PG indicate the involved CM group. CONCLUSION(S): In regular practice, most healthcare professionals and fitness trainers prefer ASTKE due to the ease of the testing procedure. Directing co-professionals on their choice of tests is challenging, whereas providing knowledge about CM use paves the way for creating customized treatment plans.

Integration of Postbiotics in Food Products through Attenuated Probiotics: A Case Study with Lactic Acid Bacteria in Bread.

The study examines the integration of postbiotics in food products through the use of attenuated probiotics, specifically lactic acid bacteria (LAB) in bread. Postbiotics, non-viable microorganisms or their metabolites, offer health benefits similar to probiotics without the risks associated with live bacteria. This research evaluates the regulatory aspects and safety of LAB in sourdough bread production, highlighting their historical and significant use in Europe before 1997. The study includes microbial quantification and Next-Generation Sequencing (NGS) to identify LAB in traditional sourdough, comparing them with historical and current EFSA Qualified Presumption of Safety (QPS) lists. Findings show that the LAB present in sourdough have been extensively and safely used in bread making, supporting their classification as non-novel foods under EU regulations. The stability and consistency of LAB metabolites in sourdough bread are also confirmed, ensuring quality and safety in each batch. The study concludes that LAB in sourdough, when inactivated through bread-making processes, are not considered novel foods, aligning with historical, scientific, and regulatory evidence.

Insights into the Versatility of Using Atomic Absorption Spectrometry in Antibacterial Research.

The ongoing development of bacterial resistance to antibiotics is a global challenge. Research in that field is thus necessary. Analytical techniques are required for such a purpose. From this perspective, the focus was on atomic absorption spectrometry (AAS). Although it is old, AAS often offers unexpected potential. Of course, this should be exploited. The aim was therefore to demonstrate the versatility of the technique in antibacterial research. This is illustrated by various examples of its practical application. AAS can be used, for example, to confirm the identity of antibacterial compounds, for purity controls, or to quantify the antibiotics in pharmaceutical preparations. The latter allowed analysis without laborious sample preparation and without interference from other excipients. In addition, AAS can help elucidate the mode of action or resistance mechanisms. In this context, quantifying the accumulation of the antibiotic drug in the cell of (resistant) bacteria appears to play an important role. The general application of AAS is not limited to metal-containing drugs, but also enables the determination of some organic chemical antibiotics. Altogether, this perspective presents a range of applications for AAS in antibacterial research, intending to raise awareness of the method and may thus contribute to the fight against resistance.

Methane Quantification Performance of the Quantitative Optical Gas Imaging (QOGI) System Using Single-Blind Controlled Release Assessment.

Quantitative optical gas imaging (QOGI) system can rapidly quantify leaks detected by optical gas imaging (OGI) cameras across the oil and gas supply chain. A comprehensive evaluation of the QOGI system's quantification capability is needed for the successful adoption of the technology. This study conducted single-blind experiments to examine the quantification performance of the FLIR QL320 QOGI system under near-field conditions at a pseudo-realistic, outdoor, controlled testing facility that mimics upstream and midstream natural gas operations. The study completed 357 individual measurements across 26 controlled releases and 71 camera positions for release rates between 0.1 kg Ch4/h and 2.9 kg Ch4/h of compressed natural gas (which accounts for more than 90% of typical component-level leaks in several production facilities). The majority (75%) of measurements were within a quantification factor of 3 (quantification error of -67% to 200%) with individual errors between -90% and 831%, which reduced to -79% to +297% when the mean of estimates of the same controlled release from multiple camera positions was considered. Performance improved with increasing release rate, using clear sky as plume background, and at wind speeds ≤1 mph relative to other measurement conditions.

Estimation of Prediction Intervals for Performance Assessment of Building Using Machine Learning.

This study utilizes artificial neural networks (ANN) to estimate prediction intervals (PI) for seismic performance assessment of buildings subjected to long-term ground motion. To address the uncertainty quantification in structural health monitoring (SHM), the quality-driven lower upper bound estimation (QD-LUBE) has been opted for global probabilistic assessment of damage at local and global levels, unlike traditional methods. A distribution-free machine learning model has been adopted for enhanced reliability in quantifying uncertainty and ensuring robustness in post-earthquake probabilistic assessments and early warning systems. The distribution-free machine learning model is capable of quantifying uncertainty with high accuracy as compared to previous methods such as the bootstrap method, etc. This research demonstrates the efficacy of the QD-LUBE method in complex seismic risk assessment scenarios, thereby contributing significant enhancement in building resilience and disaster management strategies. This study also validates the findings through fragility curve analysis, offering comprehensive insights into structural damage assessment and mitigation strategies.