AI in Measurement Science

Fingerprint Profiling of Glycans on Extracellular Vesicles via Lectin-Induced Aggregation Strategy for Precise Cancer Diagnostics

Extracellular vesicles (EVs) harbor abundant glycans that mediate various functions, such as intercellular communication and disease advancement, which play significant roles in disease progression. However, the presence of EV heterogeneity in body fluids and the complex nature of the glycan structures have posed challenges for the detection of EV glycans. In this study, we provide a streamlined method integrated, membrane-specific separation with lectin-induced aggregation strategy (MESSAGE), for multiplexed profiling of EV glycans. By leveraging a rationally designed lectin-induced aggregation strategy, the expression of EV glycans is converted to size-based signals. With the assistance learning machine algorithms, the MESSAGE strategy with high sensitivity, specificity, and simplicity can be used for early cancer diagnosis and classification, as well as monitoring cancer metastasis via 20 μL plasma sample within 2 h. Furthermore, our platform holds promise for advancing the field of EV-based liquid biopsy for clinical applications, opening new possibilities for the profiling of EV glycan signatures in various disease states.

Automated pretreatment of environmental water samples and non-targeted intelligent screening of organic compounds based on machine experiments

The complexity of environmental pollutants poses significant challenges for monitoring and analysis, especially with the emergence of numerous emerging contaminants. Traditional analysis methods rely mainly on laboratory analysis, which involves labor-intensive and time-consuming sample preparation procedures and non-target data analysis, greatly limiting the rapid detection of water organic pollutants. In this study, we designed a robot experimenter combined with GC × GC-TOFMS. By configuring self-developed automated analysis software, we established a fully automated process from sample collection to data characterization, for the analysis of organic pollutants. We validated the method with 111 organic standards compounds. The robot performed 2577 actions covering the entire workflow, from water sample collection to sample pre-treatment. The integration of mass spectrometry and related software enabled the automatic analysis of emerging hazardous contaminants, from sampling to the output of detection results. The results showed the automated process could qualitatively identify all compounds and demonstrated good linearity, low detection limits, and excellent quantitative ability within the range of 0.04-0.4 mg/L. The average recoveries of 82.89 % of the samples ranged from 70 % to 120 % (relative standard deviation (RSD) <15 %) at different spiked concentrations. This indicated that the established method could be used for non-targeted analysis of emerging contaminants in environmental water samples. We applied the method to samples from wastewater treatment plants and river sections, identifying 1,902 compounds across 26 categories, including 6 known hazardous contaminants found in all samples. The relative content of these characteristic compounds will inform whether treated wastewater meets discharge standards and aid in tracing the sources of pollutants. Therefore, the development of this fully automated machine experimental method enables real-time and online automatic analysis of organic pollutants in environmental water. The establishment of characteristic fingerprints can provide technical support for early warning and traceability of water quality.

Multispectral 3D DNA Machine Combined with Multimodal Machine Learning for Noninvasive Precise Diagnosis of Bladder Cancer

Extracellular vesicle (EV) molecular phenotyping offers enormous opportunities for cancer diagnostics. However, the majority of the associated studies adopted biomarker-based unimodal analysis to achieve cancer diagnosis, which has high false positives and low precision. Herein, we report a multimodal platform for the high-precision diagnosis of bladder cancer (BCa) through a multispectral 3D DNA machine in combination with a multimodal machine learning (ML) algorithm. The DNA machine was constructed using magnetic microparticles (MNPs) functionalized with aptamers that specifically identify the target of interest, i.e., five protein markers on bladder-cancer-derived urinary EVs (uEVs). The aptamers were hybridized with DNA-stabilized silver nanoclusters (DNA/AgNCs) and a G-quadruplex/hemin complex to form a sensing module. Such a DNA machine ensured multispectral detection of protein markers by fluorescence (FL), inductively coupled plasma mass spectrometry (ICP-MS), and UV-vis absorption (Abs). The obtained data sets then underwent uni- or multimodal ML for BCa diagnosis to compare the analytical performance. In this study, urine samples were obtained from our prospective cohort (n = 45). Our analytical results showed that the 3D DNA machine provided a detection limit of 9.2 × 103 particles mL-1 with a linear range of 4 × 104 to 5 × 107 particles mL-1 for uEVs. Moreover, the multimodal data fusion model exhibited an accuracy of 95.0%, a precision of 93.1%, and a recall rate of 93.2% on average, while those of the three types of unimodal models were no more than 91%. The elevated diagnosis precision by using the present fusion platform offers a perspective approach to diminishing the rate of misdiagnosis and overtreatment of BCa.

Advances in Machine Learning Processing of Big Data from Disease Diagnosis Sensors

Exploring accurate, noninvasive, and inexpensive disease diagnostic sensors is a critical task in the fields of chemistry, biology, and medicine. The complexity of biological systems and the explosive growth of biomarker data have driven machine learning to become a powerful tool for mining and processing big data from disease diagnosis sensors. With the development of bioinformatics and artificial intelligence (AI), machine learning models formed by data mining have been able to guide more sensitive and accurate molecular computing. This review presents an overview of big data collection approaches and fundamental machine learning algorithms and discusses recent advances in machine learning and molecular computational disease diagnostic sensors. More specifically, we highlight existing modular workflows and key opportunities and challenges for machine learning to achieve disease diagnosis through big data mining.

A Connected World: System-Level Support Through Biosensors

The goal of protecting the health of future generations is a blueprint for future biosensor design. Systems-level decision support requires that biosensors provide meaningful service to society. In this review, we summarize recent developments in cyber physical systems and biosensors connected with decision support. We identify key processes and practices that may guide the establishment of connections between user needs and biosensor engineering using an informatics approach. We call for data science and decision science to be formally connected with sensor science for understanding system complexity and realizing the ambition of biosensors-as-a-service. This review calls for a focus on quality of service early in the design process as a means to improve the meaningful value of a given biosensor. We close by noting that technology development, including biosensors and decision support systems, is a cautionary tale. The economics of scale govern the success, or failure, of any biosensor system.

Poly(ethylene oxide) Concentration Gradient-Based Microfluidic Isolation of Circulating Tumor Cells

Circulating tumor cells (CTCs) have emerged as promising circulating biomarkers for non-invasive cancer diagnosis and management. Isolation and detection of CTCs in clinical samples are challenging due to the extreme rarity and high heterogeneity of CTCs. Here, we describe a poly(ethylene oxide) (PEO) concentration gradient-based microfluidic method for rapid, label-free, highly efficient isolation of CTCs directly from whole blood samples. Stable concentration gradients of PEO were formed within the microchannel by co-injecting the side fluid (blood sample spiked with 0.025% PEO) and center fluid (0.075% PEO solution). The competition between the elastic lift force and the inertial lift force enabled size-based separation of large CTCs and small blood cells based on their distinct migration patterns. The microfluidic device could process 1 mL of blood sample in 30 min, with a separation efficiency of >90% and an enrichment ratio of >700 for tumor cells. The isolated CTCs from blood samples were enumerated by immunofluorescence staining, allowing for discrimination of breast cancer patients from healthy donors with an accuracy of 84.2%. The concentration gradient-based microfluidic separation provides a powerful tool for label-free isolation of CTCs for a wide range of clinical applications.

Multi-Pesticide Residue Analysis Method Designed for the Robot Experimenters

Robots replacing humans as the executioners is crucial work for intelligent multi-pesticide residue analysis to maximize reproducibility and throughput while minimizing the expertise required to perform the entire process. Traditional analysis methods are predicated on manual execution, so we configured our robot experimenter, automated the analytical workflow, and achieved the goal of robotics execution. Our robot experimenter with an X-Y-Z axis robotic arm was interfaced with seven modules and ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) for automated standard solution preparation, sample pre-treatment, and UPLC-MS/MS detection. An algorithm was established to make the prepared matrix-matched standard solutions meet the monitoring requirements. The strategy was demonstrated and validated for the detection of 325 pesticides in 4 typical food matrices, suggesting that the developed method is applicable for the analysis of pesticide residues in vegetables and tea as part of regulatory monitoring programs and other purposes.

Ratiometric 3D DNA Machine Combined with Machine Learning Algorithm for Ultrasensitive and High-Precision Screening of Early Urinary Diseases

Urinary extracellular vesicles (uEVs) have received considerable attention as a potential biomarker source for the diagnosis of urinary diseases. Present studies mainly focus on the discovery of biomarkers based on high-throughput proteomics, while limited efforts have been paid to applying the uEVs' biomarkers for the diagnosis of early urinary disease. Herein, we demonstrate a diagnosis protocol to realize ultrasensitive detection of uEVs and accurate classification of early urinary diseases, by combing a ratiometric three-dimensional (3D) DNA machine with machine learning (ML). The ratiometric 3D DNA machine platform is constructed by conjugating a padlock probe (PLP) containing cytosine-rich (C-rich) sequences, anchor strands, and nucleic-acid-stabilized silver nanoclusters (DNA_AgNCs) onto the magnetic nanoparticles (MNPs). The competitive binding of uEVs with the aptamer releases the walker strand, thus the ratiometric 3D DNA machine was activated to undergo an accurate amplification reaction and produce a ratiometric fluorescence signal. The present biosensor offers a detection limit of 9.9 × 10³ particles mL^-1 with a linear range of 10⁴-10⁸ particles mL^-1 for uEVs. Two ML algorithms, K-nearest neighbor (KNN) and support vector machine (SVM), were subsequently applied for analyzing the correlation between the sensing signals of uEV multibiomarkers and the clinical state. The disease diagnostic accuracy of optimal biomarker combination reaches 95% and 100% by analyzing with KNN and SVM, and the disease type classification exhibits an accuracy of 94.7% and 89.5%, respectively. Moreover, the protocol results in 100% accurate visual identification of clinical uEV samples from individuals with disease or as healthy control by a t-distributed stochastic neighbor embedding (tSNE) algorithm.