- Advanced Proteomics Techniques and Applications
- Metabolomics and Mass Spectrometry Studies
- COVID-19 Clinical Research Studies
- Mass Spectrometry Techniques and Applications
- SARS-CoV-2 and COVID-19 Research
- COVID-19 diagnosis using AI
- Machine Learning in Bioinformatics
- Long-Term Effects of COVID-19
- SARS-CoV-2 detection and testing
- Molecular Biology Techniques and Applications
- Cognitive Computing and Networks
- Machine Learning in Healthcare
- Infection Control and Ventilation
- Advanced Image and Video Retrieval Techniques
- Thyroid Cancer Diagnosis and Treatment
- Artificial Intelligence in Healthcare
- PI3K/AKT/mTOR signaling in cancer
- Bioinformatics and Genomic Networks
- Chronic Myeloid Leukemia Treatments
- Gut microbiota and health
- Thermal Regulation in Medicine
- Protein purification and stability
- Gene expression and cancer classification
- Non-Invasive Vital Sign Monitoring
- Thyroid Disorders and Treatments
Westlake University
2020-2025
University of Hong Kong
2024
Institute of Basic Medical Sciences of the Chinese Academy of Medical Sciences
2020-2023
Shenzhen University
2022-2023
Fudan University
2023
Institute for Advanced Study
2022
Shanghai Medical Information Center
2020
Beijing University of Posts and Telecommunications
2018-2019
Peking University
2015-2016
Stomatology Hospital
2016
Data-independent acquisition (DIA) mass spectrometry-based proteomics generates reproducible proteome data. The complex processing of the DIA data has led to development multiple analysis tools. In this study, we assessed performance five tools (OpenSWATH, EncyclopeDIA, Skyline, DIA-NN, and Spectronaut) using six datasets obtained from TripleTOF, Orbitrap, TimsTOF Pro instruments. By comparing identification quantification metrics examining shared unique cross-tool identifications, evaluated...
Pressure cycling technology (PCT)-assisted tissue lysis and digestion have facilitated reproducible high-throughput proteomic studies of both fresh-frozen (FF) formalin-fixed paraffin-embedded (FFPE) biopsy scale for biomarker discovery. Here, we present an improved PCT method accelerating the conventional procedures by about two-fold without sacrificing peptide yield, efficiency, peptide, protein identification. The time required processing 16 samples from tissues to peptides is reduced 6 3...
Thyroid nodules occur in about 60% of the population. A major challenge thyroid nodule diagnosis is to distinguish between follicular adenoma (FA) and carcinoma (FTC). Here, we present a comprehensive spectral library covering five types tissues. This includes 121 960 peptides 9941 protein groups. can be used quantify up 7863 proteins from tissues, also develop parallel reaction monitoring (PRM) assays for targeted quantification. Next, stratify tumours, compared proteomes 24 FA 22 FTC...
Black tooth stain is a characteristic extrinsic discoloration commonly seen on the cervical enamel following contour of gingiva. To investigate relationship between black and oral microbiota, we used 16S rRNA gene sequencing to compare microbial composition dental plaque saliva among caries-free children with without stain. Dental saliva, as well stain, were sampled from 10 15 Data analyzed using pipeline tool MOTHUR. Student's t-test was alpha diversities Mann-Whitney U test relative...
SUMMARY Severe COVID-19 patients account for most of the mortality this disease. Early detection and effective treatment severe remain major challenges. Here, we performed proteomic metabolomic profiling sera from 46 53 control individuals. We then trained a machine learning model using measurements training cohort 18 non-severe 13 patients. The correctly classified with an accuracy 93.5%, was further validated ten independent patients, seven which were classified. identified molecular...
Severity prediction of COVID-19 remains one the major clinical challenges for ongoing pandemic. Here, we have recruited a 144 patient cohort, resulting in data matrix containing 3,065 readings 124 types measurements over 52 days. A machine learning model was established to predict disease progression based on cohort consisting training, validation, and internal test sets. panel eleven routine factors constructed classifier severity prediction, achieving accuracy 98% discovery set. Validation...
Data-independent acquisition (DIA) technology for protein identification from mass spectrometry and related algorithms is developing rapidly. The spectrum-centric analysis of DIA data without the use spectra library data-dependent represents a promising direction. In this paper, we proposed an untargeted method, Dear-DIAXMBD, direct data. Dear-DIAXMBD first integrates deep variational autoencoder triplet loss to learn representations extracted fragment ion chromatograms, then uses k-means...
Abstract Serum lactate dehydrogenase (LDH) has been established as a prognostic indicator given its differential expression in COVID‐19 patients. However, the molecular mechanisms underneath remain poorly understood. In this study, 144 patients were enrolled to monitor clinical and laboratory parameters over 3 weeks. LDH was shown elevated on admission declined throughout disease course, ability classify patient severity outperformed other biochemical indicators. A threshold of 247 U/L serum...
Efficient peptide and protein identifications from data-independent acquisition mass spectrometric (DIA-MS) data typically rely on a project-specific spectral library with suitable size. Here, we describe subLib, computational strategy for optimizing the specific DIA set based comprehensive library, requiring preliminary analysis of set. Compared pan-human strategy, subLib achieved 41.2% increase in precursor 35.6% group test six colorectal tumor samples. We also applied this to 389...
RT-PCR is the primary method to diagnose COVID-19 and also used monitor disease course. This approach, however, suffers from false negatives due RNA instability poses a high risk medical practitioners. Here, we investigated potential of using serum proteomics predict viral nucleic acid positivity during COVID-19. We analyzed proteome 275 inactivated samples 54 out 144 patients shortlisted 42 regulated proteins in severe group 12 non-severe group. Using these several key clinical indexes,...
A comprehensive pan-human spectral library is critical for biomarker discovery using mass spectrometry (MS)-based proteomics. DPHL v.1, a previous built from 1,096 data-dependent acquisition (DDA) MS data of 16 human tissue types, allows quantifying 10,943 proteins. Here, we generated v.2 1,608 DDA-MS data. The included 586 acquired 18 while 1,022 files were derived v.1. thus comprises 24 sample including several cancer types (lung, breast, kidney, and prostate cancer, among others). We four...
The ongoing pandemic of the coronavirus disease 2019 (COVID-19) caused by severe acute respiratory syndrome 2 still has limited treatment options. Our understanding molecular dysregulations that occur in response to infection remains incomplete. We developed a web application COVIDpro (https://www.guomics.com/covidPro/) includes proteomics data obtained from 41 original studies conducted 32 hospitals worldwide, involving 3077 patients and covering 19 types clinical specimens, predominantly...
Batch effects are unwanted data variations that may obscure biological signals, leading to bias or errors in subsequent analyses. Effective evaluation and elimination of batch necessary for omics analysis. In order facilitate the correction effects, here we present BatchSever, an open-source R/Shiny based user-friendly interactive graphical web platform BatchServer, introduced autoComBat, a modified version ComBat, which is most widely adopted tool effect correction. BatchServer uses PVCA...
ABSTRACT Severity prediction of COVID-19 remains one the major clinical challenges for ongoing pandemic. Here, we have recruited a 144 patient cohort consisting training, validation, and internal test sets, longitudinally recorded 124 routine laboratory parameters, built machine learning model to predict disease progression based on measurements from first 12 days since onset when no became severe. A panel 11 factors, including oxygenation index, basophil counts, aspartate aminotransferase,...
The explosive growth of electronic devices brings a soaring demand for rapid electromagnetic compatibility (EMC) diagnosis. However, there is significant learning curve the electrical engineers to apply EMC knowledge. In this article, an efficient diagnosis and management methodology was proposed, which provided fast way analysis in seconds other than traditional simulation or calculation. approach organized knowledges as knowledge graph composed by interference/sensitive units, mathematical...
Abstract Here, the authors reason that complexity of medical problems and proteome science might be tackled effectively with deep learning (DL) technology. However, deployment DL for proteomics data requires acquisition sets from a large number samples. Based on success in imaging classification, thousands samples are arguably minimal input DL. Contemporary is turning high‐throughput thanks to rapid progresses sample preparation liquid chromatography mass spectrometry methods. In particular,...
A novel approach for phenotype prediction is developed data-independent acquisition (DIA) mass spectrometric (MS) data without the need peptide precursor identification using existing DIA software tools. The first step converts DIA-MS file into a new format called tensor (DIAT), which can be used convenient visualization of all ions from precursors and fragments. DIAT files fed directly deep neural network to predict phenotypes such as appearances cats, dogs, microscopic images. As proof...
Cancers are causing millions of deaths and leaving a huge clinical economic burden. High costs cancer drugs limiting their access to the growing number cases. The development more affordable alternative therapy could reach patients. As gut microbiota plays significant role in treatment cancer, microbiome-targeted has gained attention recent years. Dietary natural compounds can modulate composition while providing broader accessible medicine. Tea have been shown anti-cancer properties as well...
Abstract Colorectal cancer (CRC) is one of the most common cancers worldwide. The standard CRC chemo drug, 5-Fluorouracil (5-FU), has a poor response rate and chemoresistance, prompting need for more effective affordable treatment. In this study, we aimed to evaluate whether Prohep, novel probiotic mixture, would alleviate azoxymethane/dextran sodium sulfate (AOM/DSS)-induced colorectal tumorigenesis enhance 5-FU efficacy its mechanism. Our results suggested that Prohep showed stronger...