Deep Learning of Dual Plasma Fingerprints for High-Performance Infection Classification

Deep Learning-Enabled Rapid Metabolic Decoding of Small Extracellular Vesicles via Dual-Use Mass Spectroscopy Chip Array

The increasing focus of small extracellular vesicles (sEVs) in liquid biopsy has created a significant demand for streamlined improvements in sEV isolation methods, efficient collection of high-quality sEV data, and powerful rapid analysis of large data sets. Herein, we develop a high-throughput dual-use mass spectroscopic chip array (DUMSCA) for the rapid isolation and detection of plasma sEVs. The DUMSCA realizes more than a 50% increase in speed compared to traditional method and confirms proficiency in robust storage, reuse, high-efficiency desorption/ionization, and metabolite quantification. With the collected metabolic data matrix of sEVs, a deep learning model achieves high-performance diagnosis of Crohn's disease. Furthermore, discovered biomarkers by feature sparsification and tandem mass spectrometry experiments also exhibited remarkable performance in diagnosis. This work demonstrates the rapidity and validity of DUMSCA for disease diagnosis, enabling the diagnosis of diseases without the necessity for prior knowledge and providing a high-throughput technology for sEV-based liquid biopsy that will empower its vigorous development.

Identification of CSPG4 as a Biomarker and Therapeutic Target for Infantile Post-Hemorrhagic Hydrocephalus via Multi-Omics Analysis

Intraventricular hemorrhage in preterm neonates has become a major global health problem and is associated with a high risk of post-hemorrhagic hydrocephalus (PHH). Identifying diagnostic markers and therapeutic targets is a focal challenge in the PHH prevention and control. Here, this study applies multi-omics analyses to characterize the biochemical, proteomic, and metabolomic profiles of the cerebrospinal fluid (CSF) in clinical human cohorts to investigate disease development and recovery processes occurring due to PHH. Integrative multiomics analysis suggests that the over-representation of ferroptosis, calcium, calcium ion binding, and cell adhesion signaling pathways is associated with PHH. Bioinformatic analysis indicates that chondroitin sulfate proteoglycan 4 (CSPG4) is discovered as a CSF biomarker and positively correlated with the ventricular size and the rate of periventricular leukomalacia. Next, it is further demonstrated that these signaling pathways are dysregulated in the choroid plexus (ChP) in PHH by using in vitro cellular experiments and rat models of PHH, whereas CSPG4 silencing can suppress ferroptosis, cell adhesion function, and the intracellular flow of Ca2+. These findings broaden the understanding of the pathophysiological mechanisms of PHH and suggest that CSPG4 may be an effective therapeutic target for PHH.

Transfer Learning-Based Approach for Thickness Estimation on Optical Coherence Tomography of Varicose Veins

In-depth mechanical characterization of veins is required for promising innovations of venous substitutes and for better understanding of venous diseases. Two important physical parameters of veins are shape and thickness, which are quite challenging in soft tissues. Here, we propose the method TREE (TransfeR learning-based approach for thicknEss Estimation) to predict both the segmentation map and thickness value of the veins. This model incorporates one encoder and two decoders which are trained in a special manner to facilitate transfer learning. First, an encoder-decoder pair is trained to predict segmentation maps, then this pre-trained encoder with frozen weights is paired with a second decoder that is specifically trained to predict thickness maps. This leverages the global information gained from the segmentation model to facilitate the precise learning of the thickness model. Additionally, to improve the performance we introduce a sensitive pattern detector (SPD) module which further guides the network by extracting semantic details. The swept-source optical coherence tomography (SS-OCT) is the imaging modality for saphenous varicose vein extracted from the diseased patients. To demonstrate the performance of the model, we calculated the segmentation accuracy-0.993, mean square error in thickness (pixels) estimation-2.409 and both these metrics stand out when compared with the state-of-art methods.

Desorption Separation Ionization Mass Spectrometry (DSI-MS) for Rapid Analysis of COVID-19

During the coronavirus disease 2019 (COVID-19) pandemic, which has witnessed over 772 million confirmed cases and over 6 million deaths globally, the outbreak of COVID-19 has emerged as a significant medical challenge affecting both affluent and impoverished nations. Therefore, there is an urgent need to explore the disease mechanism and to implement rapid detection methods. To address this, we employed the desorption separation ionization (DSI) device in conjunction with a mass spectrometer for the efficient detection and screening of COVID-19 urine samples. The study encompassed patients with COVID-19, healthy controls (HC), and patients with other types of pneumonia (OP) to evaluate their urine metabolomic profiles. Subsequently, we identified the differentially expressed metabolites in the COVID-19 patients and recognized amino acid metabolism as the predominant metabolic pathway involved. Furthermore, multiple established machine learning algorithms validated the exceptional performance of the metabolites in discriminating the COVID-19 group from healthy subjects, with an area under the curve of 0.932 in the blind test set. This study collectively suggests that the small-molecule metabolites detected from urine using the DSI device allow for rapid screening of COVID-19, taking just three minutes per sample. This approach has the potential to expand our understanding of the pathophysiological mechanisms of COVID-19 and offers a way to rapidly screen patients with COVID-19 through the utilization of machine learning algorithms.

Ratiometric electrochemical OR gate assay for NSCLC-derived exosomes

Non-small cell lung cancer (NSCLC) is the most common pathological type of LC and ranks as the leading cause of cancer deaths. Circulating exosomes have emerged as a valuable biomarker for the diagnosis of NSCLC, while the performance of current electrochemical assays for exosome detection is constrained by unsatisfactory sensitivity and specificity. Here we integrated a ratiometric biosensor with an OR logic gate to form an assay for surface protein profiling of exosomes from clinical serum samples. By using the specific aptamers for recognition of clinically validated biomarkers (EpCAM and CEA), the assay enabled ultrasensitive detection of trace levels of NSCLC-derived exosomes in complex serum samples (15.1 particles μL^-1 within a linear range of 10²-10⁸ particles μL^-1). The assay outperformed the analysis of six serum biomarkers for the accurate diagnosis, staging, and prognosis of NSCLC, displaying a diagnostic sensitivity of 93.3% even at an early stage (Stage I). The assay provides an advanced tool for exosome quantification and facilitates exosome-based liquid biopsies for cancer management in clinics.