Usual Machine Learning Research Articles

BLINKRESNET: COST-EFFECTIVE EYELID LOCALIZATION FOR USE IN A BLINK REFLEXOMETER

Four keypoints corresponding to upper and lower eyelid positions may be localized in the scan videos of a blink reflexometer, an edge device. Images from prior reflexometer scans are tagged in order to train a residual neural network, dubbed BlinkResNet. In order to provide adequate demographic coverage, comprehensive data collection procedures are implemented. The architecture is thoroughly parameterized, allowing for a multidimensional space of compatible architectures to be swept through using a random search. Experimental results on a dataset of annotated reflexometer frames demonstrate significant improvement over a previous algorithm for eyelid localization. A random search identified variants of BlinkResNet best-suited for reflexometer frame annotation, with BlinkResNet-Default achieving a [Formula: see text] score of 0.966 on the validation dataset. Requirements for deployment on an edge device are met: training ease, generalizable inference performance, and low computational cost. A sweep may be conducted on any combination of hyperparameters and datasets, making it an adaptable tool for the continual development of customized machine learning architectures. The analysis of sweep results provides valuable insights into optimal model organization for specific tasks. Thus, methodical data annotation and sweeps over a parameterized machine learning routine are investigated as measures to achieve rapid, accurate inference results.

Model selection to achieve reproducible associations between resting state EEG features and autism

A concern in the field of autism electroencephalography (EEG) biomarker discovery is their lack of reproducibility. In the present study, we considered the problem of learning reproducible associations between multiple features of resting state (RS) neural activity and autism, using EEG data collected during a RS paradigm from 36 to 96 month-old children diagnosed with autism (N = 224) and neurotypical children (N = 69). Specifically, EEG spectral power and functional connectivity features were used as inputs to a regularized generalized linear model trained to predict diagnostic group (autism versus neurotypical). To evaluate our model, we proposed a procedure that quantified both the predictive generalization and reproducibility of learned associations produced by the model. When prioritizing both model predictive performance and reproducibility of associations, a highly reproducible profile of associations emerged. This profile revealed a distinct pattern of increased gamma power and connectivity in occipital and posterior midline regions associated with an autism diagnosis. Conversely, model selection based on predictive performance alone resulted in non-robust associations. Finally, we built a custom machine learning model that further empirically improved robustness of learned associations. Our results highlight the need for model selection criteria that maximize the scientific utility provided by reproducibility instead of predictive performance.

Integrating Metal-Phenolic Networks-Mediated Separation and Machine Learning-Aided Surface-Enhanced Raman Spectroscopy for Accurate Nanoplastics Quantification and Classification.

Increasing accumulation of nanoplastics across ecosystems poses a significant threat to both terrestrial and aquatic life. Surface-enhanced Raman scattering (SERS) is an emerging technique used for nanoplastics detection. However, the identification and classification of nanoplastics using SERS faces challenges regarding sensitivity and accuracy as nanoplastics are sparsely dispersed in the environment. Metal-phenolic networks (MPNs) have the potential to rapidly concentrate and separate various types and sizes of nanoplastics. SERS combined with machine learning may improve prediction accuracy. Herein, we report the integration of MPNs-mediated separation with machine learning-aided SERS methods for the accurate classification and high-precision quantification of nanoplastics, which is tailored to include the complete region of characteristic peaks across diverse nanoplastics in contrast to the traditional manual analysis of SERS spectra on a singular characteristic peak. Our customized machine learning system (e.g., outlier detection, classification, quantification) allows for the identification of detectable nanoplastics (accuracy 81.84%), accurate classification (accuracy > 97%), and sensitive quantification of various types of nanoplastics (polystyrene (PS), poly(methyl methacrylate) (PMMA), polyethylene (PE), and poly(lactic acid) (PLA)) down to ultralow concentrations (0.1 ppm) as well as accurate classification (accuracy > 92%) of nanoplastic mixtures at a subppm level. The effectiveness of this approach is substantiated by its ability to discern between different nanoplastic mixtures and detect nanoplastic samples in natural water systems.

Simultaneous Identification of Vitamins B1, B3, B5, and B6 by an Engineered Nanopore.

Vitamin Bs, a group of water-soluble compounds, are essential nutrients for almost all living organisms. However, due to their structural heterogeneity, rapid and simultaneous analysis of multiple vitamin Bs is still challenging. In this paper, it is discovered that a hetero-octameric Mycobacterium smegmatis porin A (MspA) nanopore containing a sole nickel ion-bound nitrilotriacetic acid (NTA-Ni) adapter at its pore constriction is suitable for the simultaneous sensing of different vitamin Bs, including vitamin B1 (thiamine), vitamin B3 (nicotinic acid and nicotinamide), vitamin B5 (pantothenic acid), and vitamin B6 (pyridoxine, pyridoxal, and pyridoxamine). Assisted by a custom machine learning algorithm, all seven vitamin Bs can be fully distinguished, reporting a general accuracy of 99.9%. This method was further validated in the rapid analysis of commercial cosmetics and natural food, suggesting its potential uses in food and drug administration.

Extracting typing game keystroke patterns as potential indicators of programming aptitude

This study attempted to determine whether individuals possess programming aptitude solely based on keystroke information from typing games where participants type computer programs. The participants were students enrolled in university programming courses. The results indicated that using typing speed alone as an indicator achieved an accuracy of 0.71, while employing a custom machine learning model achieved an accuracy of 0.83. Additionally, it was found that individuals with programming aptitude tended to type the enter key relatively slower compared to other keys.

Interpretable Machine Learning Models Based on SHapley Additive exPlanations for Predicting the Risk of Cerebrospinal Fluid Leakage in Lumbar Fusion Surgery.

Retrospective study. The objective of this investigation was to formulate and internally verify a customized machine learning (ML) framework for forecasting cerebrospinal fluid leakage (CSFL) in lumbar fusion surgery. This was accomplished by integrating imaging parameters and employing the SHapley Additive exPlanation (SHAP) technique to elucidate the interpretability of the model. Given the increasing incidence and surgical volume of spinal degeneration worldwide, accurate predictions of postoperative complications are urgently needed. SHAP-based interpretable ML models have not been used for CSFL risk factor analysis in lumbar fusion surgery. Clinical and imaging data were retrospectively collected from 3505 patients who underwent lumbar fusion surgery. Six distinct machine learning models were formulated: extreme gradient boosting (XGBoost), decision tree (DT), random forest (RF), support vector machine (SVM), Gaussian naive Bayes (GaussianNB), and K-nearest neighbors (KNN) models. Evaluation of model performance on the test dataset was performed using performance metrics, and the analysis was executed through the SHAP framework. CSFL was detected in 95 out of 3505 patients (2.71%). Notably, the XGBoost model exhibited outstanding accuracy in forecasting CSFLs, with high precision (0.9815), recall (0.6667), accuracy (0.8182), F1 score (0.7347), and AUC (0.7343). Additionally, through SHAP analysis, significant predictors of CSFL were identified, including ligamentum flavum thickness, zygapophysial joint degeneration grade, central spinal stenosis grade, decompression segment count, decompression mode, intervertebral height difference, Cobb angle, intervertebral height index difference, operation mode, lumbar segment lordosis angle difference, Meyerding grade of lumbar spondylolisthesis, and revision surgery. The combination of the XGBoost model with the SHAP is an effective tool for predicting the risk of CSFL during lumbar fusion surgery. Its implementation could aid clinicians in making informed decisions, potentially enhancing patient outcomes and lowering healthcare expenses. This study advocates for the adoption of this approach in clinical settings to enhance the evaluation of CSFL risk among patients undergoing lumbar fusion.

Multiparameter quantitative analyses of diagnostic cells in brain tissues from tuberous sclerosis complex.

The advent of high-dimensional imaging offers new opportunities to molecularly characterize diagnostic cells in disorders that have previously relied on histopathological definitions. One example case is found in tuberous sclerosis complex (TSC), a developmental disorder characterized by systemic growth of benign tumors. Within resected brain tissues from patients with TSC, detection of abnormally enlarged balloon cells (BCs) is pathognomonic for this disorder. Though BCs can be identified by an expert neuropathologist, little is known about the specificity and broad applicability of protein markers for these cells, complicating classification of proposed BCs identified in experimental models of this disorder. Here, we report the development of a customized machine learning pipeline (BAlloon IDENtifier; BAIDEN) that was trained to prospectively identify BCs in tissue sections using a histological stain compatible with high-dimensional cytometry. This approach was coupled to a custom 36-antibody panel and imaging mass cytometry (IMC) to explore the expression of multiple previously proposed BC marker proteins and develop a descriptor of BC features conserved across multiple tissue samples from patients with TSC. Here, we present a modular workflow encompassing BAIDEN, a custom antibody panel, a control sample microarray, and analysis pipelines-both open-source and in-house-and apply this workflow to understand the abundance, structure, and signaling activity of BCs as an example case of how high-dimensional imaging can be applied within human tissues.

Utilising InSAR for monitoring infrastructure at national scales

Interferometric SAR (InSAR) is now a widely used tool for the monitoring of infrastructure, in particular mines, dams, and oil and gas facilities. The technology is also widely used for the monitoring of construction activities such as tunnelling. With the growth of globally available and continuously updated SAR datasets, the monitoring of entire regional and national infrastructure networks is now feasible. Here we show how continuously updated, network-wide, InSAR datasets can be used to proactively manage large infrastructure networks such as roads, railways, and water pipelines. In the UK continuously updated InSAR data (derived from Sentinel-1 SAR data) is being effectively used across the entire railway network to proactively monitor potential embankment failures, large-scale landslides, and potential hazards outside of the boundary fence. A challenge of using InSAR data in this way is the huge volume of data that is available, as such custom risk metrics and machine learning algorithms are applied to highlight some of the areas at heightened risk. InSAR is also being applied to tackle the issue of water pipe leakage caused by natural ground movements, such as soil shrink-swell. In this case, regularly updated InSAR data has been used in the UK and Australia to detect areas at elevated risk of pipeline failures based on ground movement patterns, and by combining historical movement data and known failure events it is possible to refine such models. With expanding global SAR constellations proactive, nationwide, monitoring of critical assets is now possible, with the challenge being how we derive insight from these very large datasets.

Nanopore Identification of L-, D-Lactic Acids, D-Glucose and Gluconic Acid in the Serum of Human and Animals.

DL-Lactic acid and D-glucose are important human health indicators. Their aberrant levels in body fluids may indicate a variety of human pathological conditions, suggesting an urgent need of daily monitoring. However, simultaneous and rapid analysis of DL-lactic acid and D-glucose using a sole but simple sensing system has never been reported. Here, an engineered Mycobacterium smegmatis porin A (MspA) nanopore is used to simultaneously identify DL-lactic acid and D-glucose. Highly distinguishable nanopore event features are reported. Assisted with a custom machine learning algorithm, direct identification of DL-lactic acid and D-glucose is performed with human serum, demonstrating its sensing reliability against complex and heterogeneous samples. This sensing strategy is further applied in the analysis of different animal serum samples, according to which gluconicacid is further identified. The serum samples from different animals report distinguishable levels of DL-lactic acid, D-glucose and gluconicacid, suggesting its potential applications in agricultural science and breeding industry. This sensing strategy is generally direct, rapid, economic and requires only ≈µL of input serum, suitable for point of care testing (POCT) applications.

InfusedHeart: A Novel Knowledge-Infused Learning Framework for Diagnosis of Cardiovascular Events

In the undertaken study, we have used a customized dataset termed ``Cardiac-200'' and the benchmark dataset ``PhysioNet.'' which contains 1500 heartbeat acoustic event samples (without augmentation) and 1950 samples (with augmentation) heartbeat acoustic events such as normal, murmur, extrasystole, artifact, and other unlabeled heartbeat acoustic events. The primary reason for designing a customized dataset, ``cardiac-200,'' is to balance the total number of samples into categories such as normal and abnormal heartbeat acoustic events. The average duration of the recorded heartbeat acoustic events is 10-12 s. In the undertaken study, we have analyzed and evaluated various heartbeat acoustic events using audio processing libraries such as Chromagram, Chroma-cq, Chroma-short-time Fourier transform (STFT), Chroma-cqt, and Chroma-cens to extract more information from the recorded heartbeat sound signals. The noise removal process has been carried out using local binary pattern (LBP) methodology. The noise-robust heartbeat acoustic images are classified using long short-term memory (LSTM)-convolutional neural network (CNN), recurrent neural network (RNN), LSTM, Bi-LSTM, CNN, K-means Clustering, and support vector machine (SVM) methods. The obtained results have shown that the proposed InfusedHeart Framework had outclassed all the other customized machine learning and deep learning approaches such as RNN, LSTM, Bi-LSTM, CNN, K-means Clustering, and SVM-based classification methodologies. The proposed Knowledge-infused Learning Framework has achieved an accuracy of 89.36% (without augmentation), 93.38% (with augmentation), and a standard deviation of 10.64 (without augmentation), and 6.62 (with augmentation). Furthermore, the proposed framework has been tested for various signal-to-noise ratio conditions such as SignaltoNoiseRatio0, SignaltoNoiseRatio3, SignaltoNoiseRatio6, SignaltoNoiseRatio9, SignaltoNoiseRatio12, SignaltoNoiseRatio15, and SignaltoNoiseRatio18. In the end, we have shown a detailed comparison of texture and without texture approaches and have discussed future enhancements and prospective ways for future directions.

A hybrid approach with customized machine learning classifiers and multiple feature extractors for enhancing diabetic retinopathy detection

Diabetic retinopathy (DR) is a severe global issue causing blindness if untreated, affecting millions worldwide and worsening over time. Addressing this growing concern necessitates early and accurate DR identification. This study introduces a novel approach to DR detection, combining machine learning algorithms with deep feature extraction techniques. A hybrid model is proposed by stacking predictions from diverse classifiers, such as Decision Trees, Random Forests, Support Vector Machines (SVMs), and more. Three deep learning models – MobileNetV2, DenseNet121, and InceptionResNetV2 – are employed as feature extractors from retinal images. Each classifier undergoes hyperparameter tuning for optimal performance. Using the APTOS 2019 Blindness Detection dataset, including preprocessing techniques like data augmentation and standardization, this hybrid model demonstrates promising accuracy in multi-class (95.50%) and binary classification (98.36%). Notably, DenseNet121 outperforms others. The results suggest the effectiveness of this hybrid technique in early diabetic retinopathy detection, holding significant promise for improved medical intervention.

Multi-year soundscape recordings and automated call detection reveals varied impact of moonlight on calling activity of neotropical forest katydids.

Night-time light can have profound ecological effects, even when the source is natural moonlight. The impacts of light can, however, vary substantially by taxon, habitat and geographical region. We used a custom machine learning model built with the Python package Koogu to investigate the in situ effects of moonlight on the calling activity of neotropical forest katydids over multiple years. We prioritised species with calls that were commonly detected in human annotated data, enabling us to evaluate model performance. We focused on eight species of katydids that the model identified with high precision (generally greater than 0.90) and moderate-to-high recall (minimum 0.35), ensuring that detections were generally correct and that many calls were detected. These results suggest that moonlight has modest effects on the amount of calling, with the magnitude and direction of effect varying by species: half of the species showed positive effects of light and half showed negative. These findings emphasize the importance of understanding natural history for anticipating how biological communities respond to moonlight. The methods applied in this project highlight the emerging opportunities for evaluating large quantities of data with machine learning models to address ecological questions over space and time. This article is part of the theme issue 'Towards a toolkit for global insect biodiversity monitoring'.

Application of Federated Learning for Smart Agriculture System

Federated Learning (FL) presents a ground breaking approach to addressing data privacy concerns while harnessing the power of machine learning in the agricultural sector. This paper explores the application of FL for smart agriculture, examining its potential benefits and implications. FL enables collaborative model training across decentralized data sources, allowing farmers to contribute their data without compromising privacy. In smart agriculture, FL facilitates the development of customized machine learning models for tasks such as crop yield prediction, disease detection, resource optimization, and livestock management. By leveraging data from diverse geographical regions, FL models can provide localized recommendations tailored to specific farming conditions. This paper discusses the significance of FL in enabling data-driven decision-making, promoting sustainable agricultural practices, and fostering collaboration among stakeholders. Furthermore, it explores the challenges and considerations associated with implementing FL in the agricultural sector, including data heterogeneity, communication constraints, and model aggregation. Despite these challenges, FL offers immense potential for revolutionizing agriculture by empowering farmers with actionable insights while safeguarding their data privacy.

Code-Free Machine Learning Approach for EVO-ICL Vault Prediction: A Retrospective Two-Center Study.

Establishing a development environment for machine learning is difficult for medical researchers because learning to code is a major barrier. This study aimed to improve the accuracy of a postoperative vault value prediction model for implantable collamer lens (ICL) sizing using machine learning without coding experience. We used Orange data mining, a recently developed open-source, code-free machine learning tool. This study included eye-pair data from 294 patients from the B&VIIT Eye Center and 26 patients from Kim's Eye Hospital. The model was developed using OCULUS Pentacam data from the B&VIIT Eye Center and was internally evaluated through 10-fold cross-validation. External validation was performed using data from Kim's Eye Hospital. The machine learning model was successfully trained using the data collected without coding. The random forest showed mean absolute errors of 124.8 µm and 152.4 µm for the internal 10-fold cross-validation and the external validation, respectively. For high vault prediction (>750 µm), the random forest showed areas under the curve of 0.725 and 0.760 for the internal and external validation datasets, respectively. The developed model performed better than the classic statistical regression models and the Google no-code platform. Applying a no-code machine learning tool to our ICL implantation datasets showed a more accurate prediction of the postoperative vault than the classic regression and Google no-code models. Because of significant bias in measurements and surgery between clinics, the no-code development of a customized machine learning nomogram will improve the accuracy of ICL implantation.

381 Glioma Three-Dimensional Shape Predicts Underlying Genetic Mutations

INTRODUCTION: Phenotype is the detectable expression of genotype. Despite this, little is known about how the three-dimensional structure of brain tumors correlates with their genetic biomarkers. METHODS: The UCSF-PDGM dataset was filtered for histopathologically-proven gliomas. Segmentation masks of the enhancing tumor, non-enhancing/necrotic tumor, and surrounding FLAIR abnormality were obtained from each patient’s preoperative brain MRI. All tumors were evaluated for IDH mutations and 1p/19q codeletion; all grade III and IV tumors were tested for MGMT methylation status. The segmentation masks were processed to create topological features describing the tumor’s 3D shape. These features, without other clinical variables, were used in a custom machine learning pipeline to predict the presence of IDH mutations, 1p/19q codeletion, and MGMT methylation. RESULTS: After filtration, 103 of 494 gliomas tested had an IDH mutation, 29 of 494 had a 1p/19q codeletion, and 247 of 409 had methylation of MGMT. On the blinded test-subset, the machine learning model, using only features describing shape, rendered an AUROC of 0.902 (95% CI: 0.875-0.929), specificity of 90.3% (83.9%-96.6%), and sensitivity of 75.7% (68.0%-83.3%) for IDH mutation. For 1p/19q codeletion, we found an AUROC of 0.949 (0.908-0.990), specificity of 94.5% (84.3%-100%), and sensitivity of 86.6% (79.6%-93.6%). For MGMT methylation, the performance was poor with an AUROC of 0.445 (0.385-0.504). CONCLUSIONS: The three-dimensional shape of a glioma may be used to predict the presence of some key underlying genetic mutations before biopsy. Additional research is needed to validate this, improve its fidelity, and generalize it to other biomarkers.

Cybercrime Risk Found in Employee Behavior Big Data Using Semi-Supervised Machine Learning with Personality Theories

A critical worldwide problem is that ransomware cyberattacks can be costly to organizations. Moreover, accidental employee cybercrime risk can be challenging to prevent, even by leveraging advanced computer science techniques. This exploratory project used a novel cognitive computing design with detailed explanations of the action-research case-study methodology and customized machine learning (ML) techniques, supplemented by a workflow diagram. The ML techniques included language preprocessing, normalization, tokenization, keyword association analytics, learning tree analysis, credibility/reliability/validity checks, heatmaps, and scatter plots. The author analyzed over 8 GB of employee behavior big data from a multinational Fintech company global intranet. The five-factor personality theory (FFPT) from the psychology discipline was integrated into semi-supervised ML to classify retrospective employee behavior and then identify cybercrime risk. Higher levels of employee neuroticism were associated with a greater organizational cybercrime risk, corroborating the findings in empirical publications. In stark contrast to the literature, an openness to new experiences was inversely related to cybercrime risk. The other FFPT factors, conscientiousness, agreeableness, and extroversion, had no informative association with cybercrime risk. This study introduced an interdisciplinary paradigm shift for big data cognitive computing by illustrating how to integrate a proven scientific construct into ML—personality theory from the psychology discipline—to analyze human behavior using a retrospective big data collection approach that was asserted to be more efficient, reliable, and valid as compared to traditional methods like surveys or interviews.

Abstract 7090: A scaled proteomic discovery study for prostate cancer diagnostic signatures using Proteograph workflow with trapped ion mobility mass spectrometry

Abstract Background and Aims: The low cancer specificity of serum Prostate-Specific Antigen (PSA), the principal biomarker for prostate cancer detection, leads to a high frequency of unwarranted prostate biopsies. To alleviate this deficit, we executed a proteomic discovery study seeking PSA reflex signatures using the Proteograph™ Product Suite, a multi-nanoparticle-based deep plasma/serum proteomics workflow (Seer, Inc.), with timsTOF Pro mass spectrometry (Bruker) to interrogate over 900 serum specimens from patients that underwent prostate biopsy due to elevated PSA and/or abnormal digital rectal exam. Methods: Our study followed rigorous design principles including uniform collection, randomization and blinding of specimens, all of which were processed with the Proteograph Product Suite. Liquid chromatography-mass spectrometry (LC-MS) analyses leveraged the Bruker timsTOF Pro MS platform utilizing 30-minute reversed-phase chromatography and a label-free dia-PASEF (data independent acquisition - parallel accumulation serial fragmentation) data acquisition method. Peptides deriving from a chosen subset of specimens designed to maximize peptide diversity were pooled, fractionated and analyzed using DDA (data-dependent acquisition)-PASEF to build a spectral reference library. Results: The DIA-NN algorithm was employed through the cloud-based Proteograph Analysis Suite (PAS) to search LC-MS data. Leveraging our study-specific spectral library proteomic depth achieved was approximately 3600 protein groups identified (median) per patient serum specimen and more than 4400 in at least 25% of samples across the study. Customized machine learning workflows and the SeerML pipeline were used to identify and assess diagnostic power of new proteomic signatures. Conclusions: Our data demonstrate the remarkable proteomic depth achievable in a scaled patient serum specimen discovery study using a combination of Proteograph and timsTOF platforms. This significant effort revealed serum proteomic signatures with increased diagnostic performance over that provided by PSA blood measurement and traditional patient-associated metadata alone, providing a foundation for future clinical tests aimed to reduce the current high frequency of unnecessary prostate biopsy. Citation Format: Mark Flory, Matthew Chang, Jessie May Cartier, Jane Lange, Travis Moore, James McGann, Ryan Kopp, Michael Liss, Robin Leach, Brenna Albracht, James Dai, Michael Krawitzky, Eltaher Elgierari, Jessica Chu, Paul Pease, Max Mahoney, Shadi Ferdosi, Ryan Benz, Khatereh Motamedchaboki, Asim Siddiqui, Mahdi Zamanighomi, Amir Alavi, Harendra Guturu, Daniel Hornburg, Serafim Batzoglou. A scaled proteomic discovery study for prostate cancer diagnostic signatures using Proteograph workflow with trapped ion mobility mass spectrometry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7090.

Deciphering silver nanoparticles perturbation effects and risks for soil enzymes worldwide: Insights from machine learning and soil property integration

Globally extensive research into how silver nanoparticles (AgNPs) affect enzyme activity in soils with differing properties has been limited by cost-prohibitive sampling. In this study, customized machine learning (ML) was used to extract data patterns from complex research, with a hit rate of Random Forest > Multiple Imputation by Chained Equations > Decision Tree > K-Nearest Neighbors. Results showed that soil properties played a pivotal role in determining AgNPs’ effect on soil enzymes, with the order being pH > organic matter (OM) > soil texture ≈ cation exchange capacity (CEC). Notably, soil enzyme activity was more sensitive to AgNPs in acidic soil (pH < 5.5), while elevated OM content (>1.9 %) attenuated AgNPs toxicity. Compared to soil acidification, reducing soil OM content is more detrimental in exacerbating AgNPs’ toxicity and it emerged that clay particles were deemed effective in curbing their toxicity. Meanwhile sand particles played a very different role, and a sandy soil sample at > 40 % of the water holding capacity (WHC), amplified the toxicity of AgNPs. Perturbation mapping of how soil texture alters enzyme activity under AgNPs exposure was generated, where soils with sand (45–65 %), silt (< 22 %), and clay (35–55 %) exhibited even higher probability of positive effects of AgNPs. The average calculation results indicate the sandy clay loam (75.6 %), clay (74.8 %), silt clay (65.8 %), and sandy clay (55.9 %) texture soil demonstrate less AgNPs inhibition effect. The results herein advance the prediction of the effect of AgNPs on soil enzymes globally and determine the soil types that are more sensitive to AgNPs worldwide.

Nanopore analysis of cis-diols in fruits

Natural fruits contain a large variety of cis-diols. However, due to the lack of a high-resolution sensor that can simultaneously identify all cis-diols without a need of complex sample pretreatment, direct and rapid analysis of fruits in a hand-held device has never been previously reported. Nanopore, a versatile single molecule sensor, can be specially engineered to perform this task. A hetero-octameric Mycobacterium smegmatis porin A (MspA) nanopore modified with a sole phenylboronic acid (PBA) adapter is prepared. This engineered MspA accurately recognizes 1,2-diphenols, alditols, α-hydroxy acids and saccharides in prune, grape, lemon, different varieties of kiwifruits and commercial juice products. Assisted with a custom machine learning program, an accuracy of 99.3% is reported and the sample pretreatment is significantly simplified. Enantiomers such as DL-malic acids can also be directly identified, enabling sensing of synthetic food additives. Though demonstrated with fruits, these results suggest wide applications of nanopore in food and drug administration uses.

Nanopore analysis of salvianolic acids in herbal medicines

Natural herbs, which contain pharmacologically active compounds, have been used historically as medicines. Conventionally, the analysis of chemical components in herbal medicines requires time-consuming sample separation and state-of-the-art analytical instruments. Nanopore, a versatile single molecule sensor, might be suitable to identify bioactive compounds in natural herbs. Here, a phenylboronic acid appended Mycobacterium smegmatis porin A (MspA) nanopore is used as a sensor for herbal medicines. A variety of bioactive compounds based on salvianolic acids, including caffeic acid, protocatechuic acid, protocatechualdehyde, salvianic acid A, rosmarinic acid, lithospermic acid, salvianolic acid A and salvianolic acid B are identified. Using a custom machine learning algorithm, analyte identification is performed with an accuracy of 99.0%. This sensing principle is further used with natural herbs such as Salvia miltiorrhiza, Rosemary and Prunella vulgaris. No complex sample separation or purification is required and the sensing device is highly portable.