Adaptive Sampling Approach Research Articles

Adaptive importance sampling approach for structural time-variant reliability analysis

A novel sampling approach, called adaptive importance sampling (AIS), is proposed to efficiently perform time-variant reliability analysis. In practice, structures are generally subject to time-variant deterioration processes and external loads, and the Time-variant Failure Probability Function (TFPF), which is the failure probability as a function of time, is a critical quantity of interest in engineering applications. The proposed approach leverages an adaptive strategy and an optimal combination algorithm to further improve the accuracy and efficiency of TFPF estimation using the importance sampling approach. The adaptive strategy is to seek for the best setting of importance sampling components to iteratively obtain estimator components of the TFPF. The optimal combination algorithm is to collect all these adaptive estimator components to form an overall estimator by its coefficient of variation (C.o.V.). The proposed approach outperforms traditional importance sampling methods in the sense that it ensures the convergence with minimal computational cost, specifically the C.o.V. of the TFPF estimator is below a predetermined threshold over the entire time domain. Therefore, the proposed approach offers an extension to traditional importance sampling methods for time-variant reliability assessment. Numerical examples are provided to demonstrate the effectiveness of the proposed approach in accurately estimating the TFPF of structures subjected to time-variant loads and deterioration processes.

A scalable adaptive sampling approach for surrogate modeling of rigid pavements using machine learning

Rigid pavement design is a high-dimensional optimization problem, involving several variables and design considerations. The existing machine learning (ML) design models are either low-dimensional or less accurate than computational simulations. Surrogate modeling is a powerful tool for approximating the results of high-fidelity computational simulations, which is commonly used to approximate the results of these simulations by sampling their solutions to train models. However, conventional and adaptive sampling methods face the challenge of “curse of dimensionality”, due to the exponential increase in required sample size and computations with the number of variables. To overcome this, we propose a scalable adaptive sampling (SAS) method that uses random samples for testing the model's performance but generates new training samples as a subset of a full factorial design of experiments (DoE). The factorial level increases with each iteration, allowing the algorithm to sample training data at a progressively finer scale, and updating the ML models each time adaptively. The proposed method was tested on several 2D benchmark examples, as well as practical pavement design problems. Our results demonstrate that the proposed technique can develop accurate surrogate models for both low- and high-dimensional inference spaces. For a 4D inference space, the surrogate models derived from the proposed method had an order of magnitude lower error than those derived from “one-shot” conventional sampling with the same sample size. For a 6D inference space, the sample size required by the proposed method was only 5 % of that required by conventional sampling for comparable performance.

Adaptive sampling by citizen scientists improves species distribution model performance: A simulation study

Abstract Volunteer recorders generate large amounts of biodiversity data through citizen science which is used in conservation planning and policy decision‐making. Unstructured sampling, where the volunteer can record what they want, where they want, leads to spatial unevenness in these data. While there are many statistical techniques to account for the resulting biases, it may be possible to improve datasets by directing a subset of recorders to sample in the most informative locations, known as adaptive sampling. We investigated the potential for adaptive sampling to improve the performance of species distribution models built on citizen science data using simulated ecological communities. We simulated ecological assemblages across Great Britain based on current butterfly data and modelled the distributions of each species. We then simulated the sampling of new data based on five adaptive sampling methods (one empirical method based simply on gap‐filling, and four model‐based methods using various measures from the model outputs) and one non‐adaptive method (a method in which recording continued in the current pattern), and re‐ran the species distribution models. In these, we also varied the rate of recording effort that was distributed according to adaptive sampling. The model predictions using the original and adaptively sampled data were compared to true species distributions to evaluate the performance of each method. We found that all adaptive sampling approaches improved model performance, with greatest improvement for model‐based approaches compared to the empirical sampling method (i.e. simple gap‐filling). All four model‐based adaptive sampling approaches provided similar benefits for model outputs. Improvements in model performance were greatest when the amount of adaptive sampling changed from no uptake to 1% uptake, indicating that only a small amount of change in recorder behaviour is needed to improve model performance. Directing volunteer recorders to places where records are most needed, based on information from model outputs, can improve species distribution models built on citizen science data, even with minimal uptake of suggested locations. Our results therefore suggest that adaptive sampling by recorders could be beneficial for real‐world citizen science datasets.

Generating in silico protein structural ensembles that agree with solution-based experiments: An HDX-MS-steered adaptive sampling approach for molecular dynamics

Generating in silico protein structural ensembles that agree with solution-based experiments: An HDX-MS-steered adaptive sampling approach for molecular dynamics

A fault diagnosis method for nuclear power plants rotating machinery based on deep learning under imbalanced samples

Rotating machinery is a critical equipment widely used in nuclear power plants (NPPs). Effective fault diagnosis technology can provide reliable operation and maintenance support for rotating machinery. Data-driven intelligent fault diagnosis technology has attracted much attention in recent years. However, in practical situations, the available fault data is limited, and the imbalanced samples will make the intelligent fault diagnosis model prone to problems such as poor generalization performance and low diagnostic accuracy. Therefore, a fault diagnosis method based on deep learning is proposed to alleviate the impact of imbalanced samples on fault diagnosis of rotating machinery. First, the adaptive synthetic sampling (ADASYN) approach is used to synthesize the multi-channel vibration data to expand the unbalanced samples. Subsequently, ensemble empirical mode decomposition (EEMD) and continuous wavelet transform (CWT) are used to convert the vibration data into time–frequency images to highlight the fault features of the samples. Then, the deep residual neural network is constructed to extract the features of mixed samples and implement fault diagnosis. Fault simulation experiments are carried out based on motors and bearings, and various imbalance degree datasets are formed to provide data support for verifying the effectiveness of the proposed method. The proposed method can achieve good diagnostic performance under various degrees of imbalanced samples. In addition, the comparison results with other methods show that the proposed method has the best comprehensive performance, demonstrating the potential application value in the fault diagnosis of NPPs rotating machinery.

WashRing: An Energy-Efficient and Highly Accurate Handwashing Monitoring System via Smart Ring

The outbreak of COVID-19 has greatly changed everyone's lifestyle all over the world. One of the best ways to prevent the spread of infections is by washing hands properly. Although a number of hand hygiene monitoring systems have been proposed, they either cannot achieve high accuracy in practice or work only in limited environments such as hospitals. Therefore, a ubiquitous, energy-efficient and highly accurate hand hygiene monitoring system is still lacking. In this paper, we present <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">WashRing</i> —the first smart ring-based handwashing monitoring system. In WashRing, we design a Partially Observable Markov Decision Process (POMDP) based adaptive sampling approach to achieve high energy efficiency. Then, we design an automatic feature extraction scheme based on wavelet scattering and a CNN-LSTM neural network to achieve fine-grained gesture recognition. Finally, we model the handwashing gesture classification as a few-shot learning problem to mitigate the burden of collecting extensive data from five fingers. We collect data from 25 subjects over 2 months and evaluate the system performance on both commercial OURA ring and customized ring. Evaluation results show that WashRing achieves 97.8% accuracy which is 10.2%–15.9% higher than state-of-the-arts. Our adaptive sampling approach reduces energy consumption by 64.2% compared to fixed duty cycle sampling strategies.

Investigating adaptive sampling strategies for optimal building energy performance using artificial neural networks and kriging surrogate models

Recent studies highlight surrogate models as efficient alternatives to energy simulation tools, reducing computational burden. However, accurate metamodels often require extensive evaluations, leading to high computational costs. Adaptive sampling approaches offer a promising solution by constructing accurate surrogate models at a lower computational expense compared to static sampling. The selection of an appropriate adaptive sampling criterion is crucial in sequential sampling strategies, as it determines how new sample points are chosen based on previous iterations. However, concerns remain regarding the impact of the sampling infill criterion on surrogate model accuracy. Additionally, a lack of studies exists comparing artificial neural networks (ANN) and Kriging models in terms of potential computational improvements through sequential sampling. This study aims to explore the potential of adaptive sampling strategies to estimate optimal building energy consumption within a reasonable computational budget, thus reducing the number of energy simulation model runs required. Furthermore, a comprehensive comparison between sequential sampling techniques employing ANN-based and Kriging-based surrogate models is conducted. Four different sampling criteria are evaluated to understand the relationship between sampling strategies and accuracy. The findings contribute to a more efficient resolution of building optimization problems. The results indicate that the choice of a sampling strategy depends on the model's requirements and the specific problem. Each strategy has distinct strengths and weaknesses, necessitating careful consideration of the model's needs. Selecting the appropriate sampling strategy improves model accuracy and reduces training time. Adaptive sampling methods yield significant improvements, saving 25 %–68 % of simulation samples compared to static approaches.

Oversampling methods for machine learning model data training to improve model capabilities to predict the presence of Escherichia coli MG1655 in spinach wash water.

We assessed the efficacy of oversampling techniques to enhance machine learning model performance in predicting Escherichia coli MG1655 presence in spinach wash water. Three oversampling methods were applied to balance two datasets, forming the basis for training random forest (RF), support vector machines (SVMs), and binomial logistic regression (BLR) models. Data underwent method-specific centering and standardization, with outliers replaced by feature-specific means in training datasets. Testing occurred without these preprocessing steps. Model hyperparameters were optimized using a subset of testing data via 10-fold cross-validation. Models were trained on full datasets and tested on newly acquired spinach wash water samples. Synthetic Minority Oversampling Technique (SMOTE) and Adaptive Synthetic Sampling approach (ADASYN) achieved strong results, with SMOTE RF reaching an accuracy of 90.0%, sensitivity of 93.8%, specificity of 87.5%, and an area under the curve (AUC) of 98.2% (without data preprocessing) and ADASYN achieving 86.55% accuracy, 87.5% sensitivity, 83.3% specificity, and a 92.4% AUC. SMOTE and ADASYN significantly improved (p<0.05) SVM and RF models, compared to their non-oversampled counterparts without preprocessing. Data preprocessing had a mixed impact, improving (p<0.05) the accuracy and specificity of the BLR model but decreasing the accuracy and specificity (p<0.05) of the SVM and RF models. The most influential physiochemical feature for E. coli detection in wash water was water conductivity, ranging from 7.9 to 196.2 µS. Following closely was water turbidity, ranging from 2.97 to 72.35 NTU within this study.

Logistic Regression Ensemble Classifier for Intrusion Detection System in Internet of Things

The Internet of Things (IoT) is a powerful technology that connect its users worldwide with everyday objects without any human interference. On the contrary, the utilization of IoT infrastructure in different fields such as smart homes, healthcare and transportation also raises potential risks of attacks and anomalies caused through node security breaches. Therefore, an Intrusion Detection System (IDS) must be developed to largely scale up the security of IoT technologies. This paper proposes a Logistic Regression based Ensemble Classifier (LREC) for effective IDS implementation. The LREC combines AdaBoost and Random Forest (RF) to develop an effective classifier using the iterative ensemble approach. The issue of data imbalance is avoided by using the adaptive synthetic sampling (ADASYN) approach. Further, inappropriate features are eliminated using recursive feature elimination (RFE). There are two different datasets, namely BoT-IoT and TON-IoT, for analyzing the proposed RFE-LREC method. The RFE-LREC is analyzed on the basis of accuracy, recall, precision, F1-score, false alarm rate (FAR), receiver operating characteristic (ROC) curve, true negative rate (TNR) and Matthews correlation coefficient (MCC). The existing researches, namely NetFlow-based feature set, TL-IDS and LSTM, are used to compare with the RFE-LREC. The classification accuracy of RFE-LREC for the BoT-IoT dataset is 99.99%, which is higher when compared to those of TL-IDS and LSTM.

Domain Adaptive Sampling for Cross-Domain Point Cloud Recognition

Point cloud recognition has recently gained increasing research interest due to the huge potential in real-world applications such as autonomous driving, robotics, etc. However, the point clouds of similar objects often exhibit notable geometric variations due to the difference in capturing devices or environmental changes. This leads to significant performance degradation when the learned point cloud recognition model is applied to a new scenario, which is also known as the domain adaptation issue. In this work, we propose a new unsupervised domain adaptation approach for point cloud recognition via domain adaptive sampling (DAS). In particular, we propose a two-level sampling strategy of point level and instance level to improve the cross-domain recognition ability of the model. First, we propose a domain adaptive point sampling (DAPS) strategy to enhance the domain-invariant representation of point clouds by progressively focusing on representative points in each point cloud based on geometric consistency. Then, we further propose an instance-level domain adaptive cloud sampling (DACS) strategy to learn target-specific information based on a self-paced learning paradigm, where we select a set of pseudo-labeled target point clouds to train our designed light-weighted adapters without modifying the learned domain-invariant representation. We validate our domain adaptive sampling approach on the benchmark datasets PointDA-10 and GraspNetPC-10, where our method achieves new state-of-the-art performance.

The use of class imbalanced learning methods on ULSAM data to predict the case–control status in genome-wide association studies

Machine learning (ML) methods for uncovering single nucleotide polymorphisms (SNPs) in genome-wide association study (GWAS) data that can be used to predict disease outcomes are becoming increasingly used in genetic research. Two issues with the use of ML models are finding the correct method for dealing with imbalanced data and data training. This article compares three ML models to identify SNPs that predict type 2 diabetes (T2D) status using the Support vector machine SMOTE (SVM SMOTE), The Adaptive Synthetic Sampling Approach (ADASYN), Random under sampling (RUS) on GWAS data from elderly male participants (165 cases and 951 controls) from the Uppsala Longitudinal Study of Adult Men (ULSAM). It was also applied to SNPs selected by the SMOTE, SVM SMOTE, ADASYN, and RUS clumping method. The analysis was performed using three different ML models: (i) support vector machine (SVM), (ii) multilayer perceptron (MLP) and (iii) random forests (RF). The accuracy of the case–control classification was compared between these three methods. The best classification algorithm was a combination of MLP and SMOTE (97% accuracy). Both RF and SVM achieved good accuracy results of over 90%. Overall, methods used against unbalanced data, all three ML algorithms were found to improve prediction accuracy.

Radiomics from dual-energy CT-derived iodine maps predict lymph node metastasis in head and neck squamous cell carcinoma.

To develop and validate an iodine maps-based radiomics nomogram for preoperatively predicting cervical lymph node metastasis (LNM) in head and neck squamous cell carcinoma (HNSCC). A total of 278 patients who pathologically confirmed as HNSCC were retrospectively recruited from two medical centers between June 2012 and July 2022. The training set (n = 152) and internal set (n = 67) were randomly selected from medical center A, and the patients from medical center B were enrolled as the external set (n = 69). The minority group in the training set was balanced by the adaptive synthetic sampling (ADASYN) approach. Radiomics features were extracted from dual-energy CT-derived iodine maps at arterial phase (AP) and venous phase (VP), respectively. Three radiomics signatures were constructed to predict the LNM by using a random forest algorithm. The independent clinical predictors for LNM were identified by multivariate analysis and combined with radiomics signatures to establish a radiomic-clinical nomogram. The performance of radiomic-clinical nomogram was evaluated with respect to its discrimination and clinical usefulness. The AP-VP-incorporated radiomics model exhibited a great predictive performance for LNM prediction with an area under curve (AUC) of 0.885 (95% CI, 0.836-0.933) in ADASYN-training set and confirmed in all validation sets. The nomogram that incorporated AP-VP radiomics signatures, CT-reported LN status, and histological grades yielded AUCs of 0.920 (95% CI, 0.881-0.959) in ADASYN-training set, 0.858 (95% CI, 0.771-0.944) in internal validation, and 0.849 (95% CI, 0.752-0.946) in external validation, with good calibration in all cohorts (p > 0.05). Decision curve analyses indicated the nomogram was clinically useful. In addition, the predictive performance of clinical-radiomics nomogram was also validation in combing cohorts. Stratified analysis confirmed the stability of nomogram, particularly in group negative for CT-reported LNM. Clinical-radiomics nomogram based on iodine maps exhibited promising performance in predicting LNM and providing valuable information for making individualized therapy decisions.

Hybrid resampling and weighted majority voting for multi-class anomaly detection on imbalanced malware and network traffic data

In a large skewed dataset, the data imbalance is severe and the classifier's accuracy is biased towards the majority class. Insufficient data makes it challenging for the classifier to learn the feature of the minority classes. Moreover, some existing techniques of binary classification cannot directly apply to multi-classification. Plain oversampling may generate redundant and irrelevant data for the minority classes while undersampling may eliminate too many features for the majority class. This paper proposes a combination of constrained undersampling, oversampling, noise cleaning, and a weighted majority voting classifier (WMVC) to detect multi-classification obfuscated malware via memory and network traffic anomalies. As a constraint on undersampling and oversampling, the proposed framework divides the total sample number by the class number to obtain the average class size. According to the constraint, one or more majority classes are down-sampled using Random Undersampling, while all the minority classes are up-sampled using Adaptive Synthetic Sampling Approach, which is followed by Tomek Link to remove the noisy data. Then a weighted majority voting classifier aggregated tree-based ensemble algorithms is designed and compared to the XGBoost and Convolutional Neural Network (CNN) classifiers and six state-of-the-art ensemble algorithms. The comparison results show that the performance of the classifiers on balanced data outperforms those on imbalanced data, and the WMVC outperforms XGBoost, CNN, and the six other ensemble algorithms. Our approach can alleviate the classifier's bias towards the majority class while improving its performance for the difficult minority class.

Expert-demonstration-augmented reinforcement learning for lane-change-aware eco-driving traversing consecutive traffic lights

Eco-driving methods incorporating lateral motion exhibit enhanced energy-saving prospects in multi-lane traffic contexts, yet the randomly distributed obstructing vehicles and sparse traffic lights pose challenges in assessing the long-term value of instantaneous actions, impeding further improvement in energy efficiency. In response to this issue, a deep reinforcement learning (DRL)-based eco-driving method is proposed and augmented with the expert demonstration mechanism. Specifically, a Markov decision process matching with the target eco-driving scenario is systematically constructed, with which, the formulated DRL algorithm, parametrized soft actor-critic (PSAC), is trained to realize the integrated optimization of speed planning and lane-changing maneuver. To promote the training performance of PSAC under sparse rewards concerning traffic lights, an expert eco-driving model and an adaptive sampling approach are incorporated to constitute the expert demonstration mechanism. Simulation results highlight the superior performance of the proposed DRL-based eco-driving method and its training mechanism. Compared with the performance of the PSAC with a pure exploration-based training mechanism, the expert demonstration mechanism promotes the training efficiency and cumulated rewards of PSAC by about 60 % and 21.89 % respectively in the training phase, while in the test phase, a further reduction of 4.23 % benchmarked on a rule-based method is achieved in fuel consumption.

Deadline-aware misinformation prevention in social networks with time-decaying influence

A misinformation prevention problem is essential in social networks since misinformation could greatly mislead people and interfere societal, economical, or even political circumstances. Traditional diffusion models for describing information spread assume that active individuals have only one trial to activate their friends. This may oversimplify the mechanism of the influence between two individuals in reality, i.e., an individual can have more than one chance to influence its neighbors. In addition, the influence between nodes is considered as a constant, although it is inherently dynamic in nature. This paper addresses the deadline-aware misinformation prevention problem considering a more realistic diffusion model, i.e., competitive dynamic independent cascade model (CDIC), which allows multiple activation trials with time-decaying influence power. We further show that the CDIC preserves the monotone and submodular properties. The deadline-aware misinformation prevention problem aims to find some influential truth spreaders such that the expected number of nodes to be saved or protected from misinformation is maximized before a given deadline. An adaptive sampling approach (ASA) is proposed under the CDIC model for solving the problem, which outperforms the greedy approach with lazy evaluations by several orders of magnitude in terms of the computational efficiency through a series of simulation on real datasets.

Adaptive sampling for nanopore direct RNA-sequencing.

Nanopore long-read sequencing enables real-time monitoring and controlling of individual nanopores. This allows us to enrich or deplete specific sequences in DNA sequencing in a process called "adaptive sampling." So far, adaptive sampling (AS) was not applicable to the direct sequencing of RNA. Here, we show that AS is feasible and useful for direct RNA sequencing (DRS), which has its specific technical and biological challenges. Using a well-controlled in vitro transcript-based model system, we identify essential characteristics and parameter settings for AS in DRS, as the superior performance of depletion over enrichment. Here, the efficiency of depletion is close to the theoretical maximum. Additionally, we demonstrate that AS efficiently depletes specific transcripts in transcriptome-wide sequencing applications. Specifically, we applied our AS approach to poly(A)-enriched RNA samples from human-induced pluripotent stem cell-derived cardiomyocytes and mouse whole heart tissue and show efficient 2.5- to 2.8-fold depletion of highly abundant mitochondrial-encoded transcripts. Finally, we characterize depletion and enrichment performance for complex transcriptome subsets, that is, at the level of the entire Chromosome 11, proving the general applicability of direct RNA AS. Our analyses provide evidence that AS is especially useful to enable the detection of lowly expressed transcripts and reduce the sequencing of highly abundant disturbing transcripts.

Binding-and-Folding Recognition of an Intrinsically Disordered Protein Using Online Learning Molecular Dynamics.

Intrinsically disordered proteins participate in many biological processes by folding upon binding to other proteins. However, coupled folding and binding processes are not well understood from an atomistic point of view. One of the main questions is whether folding occurs prior to or after binding. Here we use a novel, unbiased, high-throughput adaptive sampling approach to reconstruct the binding and folding between the disordered transactivation domain of c-Myb and the KIX domain of the CREB-binding protein. The reconstructed long-term dynamical process highlights the binding of a short stretch of amino acids on c-Myb as a folded α-helix. Leucine residues, especially Leu298-Leu302, establish initial native contacts that prime the binding and folding of the rest of the peptide, with a mixture of conformational selection on the N-terminal region with an induced fit of the C-terminal.

Multi-fidelity surrogate-based optimization for microfluidic concentration gradient generator design

PurposeThis paper aims to present a multi-fidelity surrogate-based optimization (MFSBO) method for computationally accurate and efficient design of microfluidic concentration gradient generators (µCGGs).Design/methodology/approachCokriging-based multi-fidelity surrogate model (MFSM) is constructed to combine data with varying fidelities and computational costs to accelerate the optimization process and improve design accuracy. An adaptive sampling approach based on parallel infill of multiple low-fidelity (LF) samples without notably adding computation burden is developed. The proposed optimization framework is compared with a surrogate-based optimization (SBO) method that relies on data from a single source, and a conventional multi-fidelity adaptive sampling and optimization method in terms of the convergence rate and design accuracy.FindingsThe results demonstrate that proposed MFSBO method allows faster convergence and better designs than SBO for all case studies with 49% more reduction in the objective function value on average. It is also found that parallel infill (MFSBO-4) with four LF samples, enables more robust, efficient and accurate designs than conventional multi-fidelity infill (MFSBO-1) that only adopts one LF sample during each iteration for more complex optimization problems.Originality/valueA MFSM based on cokriging method is constructed to utilize data with varying fidelities, accuracies and computational costs for µCGG design. A parallel infill strategy based on multiple infill criteria is developed to accelerate the convergence and improve the design accuracy of optimization. The proposed methodology is proved to be a feasible method for µCGG design and its computational efficiency is verified.

GAS-AU: an average uncertainty-based general adaptive sampling approach

GAS-AU: an average uncertainty-based general adaptive sampling approach

Benchmarking microbial DNA enrichment protocols from human intestinal biopsies.

Shotgun metagenomic sequencing is a powerful tool for studying bacterial communities in their natural habitats or sites of infection, without the need for cultivation. However, low microbial signals in metagenomic sequencing can be overwhelmed by host DNA contamination, resulting in decreased sensitivity for microbial read detection. Several commercial kits and other methods have been developed to enrich bacterial sequences; however, these assays have not been tested extensively for human intestinal tissues yet. Therefore, the objective of this study was to assess the effectiveness of various wet-lab and software-based approaches for depleting host DNA from microbiome samples. Four different microbiome DNA enrichment methods, namely the NEBNext Microbiome DNA Enrichment kit, Molzym Ultra-Deep Microbiome Prep, QIAamp DNA Microbiome kit, and Zymo HostZERO microbial DNA kit, were evaluated, along with a software-controlled adaptive sampling (AS) approach by Oxford Nanopore Technologies (ONT) providing microbial signal enrichment by aborting unwanted host DNA sequencing. The NEBNext and QIAamp kits proved to be effective in shotgun metagenomic sequencing studies, as they efficiently reduced host DNA contamination, resulting in 24% and 28% bacterial DNA sequences, respectively, compared to <1% in the AllPrep controls. Additional optimization steps using further detergents and bead-beating steps improved the efficacy of less efficient protocols but not of the QIAamp kit. In contrast, ONT AS increased the overall number of bacterial reads resulting in a better bacterial metagenomic assembly with more bacterial contigs with greater completeness compared to non-AS approaches. Additionally, AS also allowed for the recovery of antimicrobial resistance markers and the identification of plasmids, demonstrating the potential utility of AS for targeted sequencing of microbial signals in complex samples with high amounts of host DNA. However, ONT AS resulted in relevant shifts in the observed bacterial abundance, including 2 to 5 times more Escherichia coli reads. Furthermore, a modest enrichment of Bacteroides fragilis and Bacteroides thetaiotaomicron was also observed with AS. Overall, this study provides insight into the efficacy and limitations of various methods for reducing host DNA contamination in human intestinal samples to improve the utility of metagenomic sequencing.