TMT Proteomics: Experimental Design, Labeling, and Data Analysis

Tandem mass tag (TMT) proteomics is a multiplexed quantitative mass spectrometry approach that labels peptides from multiple biological samples with isobaric chemical tags, allowing simultaneous relative quantification across experimental conditions within a single LC-MS/MS run. This article provides practical guidance for researchers, students, analysts, and life-science professionals who need to design, execute, and interpret TMT-based quantitative proteomics experiments. The scope covers experimental design decisions, labeling chemistry considerations, quality control measures, data analysis pipelines, and common pitfalls that affect data quality and biological interpretation.

At a Glance

TMT reagents enable parallel quantification of up to 16 or more samples in a single mass spectrometry experiment. The core workflow involves protein extraction, digestion into peptides, chemical labeling with TMT tags, sample pooling, fractionation, LC-MS/MS analysis, and computational processing. The table below summarizes key decision points across the experimental pipeline.

Scope and Context of TMT-Based Quantitative Proteomics

TMT-based proteomics has become a standard approach for comparing protein abundances across multiple conditions in a single experiment. The method relies on isobaric tags that have identical total mass but produce distinct reporter ions upon fragmentation in the mass spectrometer. This design allows samples to be combined before analysis, reducing instrument time and technical variability compared to label-free approaches.

The practical utility of TMT extends across diverse biological questions. Researchers have applied TMT labeling to study autophagy perturbations, where quantitative proteomics reveals changes in protein abundance under starvation or infection conditions [8]. The approach also supports phosphoproteome profiling, with modular pipelines available for both TMT-based data-dependent acquisition and label-free data-independent acquisition workflows [9]. In clinical research, TMT labeling has enabled large-scale plasma proteomics studies, such as the Vanderbilt Memory and Aging Project, which identified plasma proteins predictive of cognitive decline over a 9-year follow-up period using 23 16-plex TMT batches [10].

For plant research, optimized TMT protocols allow quantification of more than 8000 proteins from Arabidopsis samples across multiple conditions, requiring only 0.1 mg protein per sample [11]. Single-cell applications have also emerged, with the SCoPE2 protocol using isobaric carriers to enhance peptide sequence identification and enabling quantification of over 1000 proteins per cell [6].

The choice between TMT and alternative quantification strategies depends on experimental goals. Label-free approaches offer flexibility in sample number but require more instrument time and are more susceptible to run-to-run variability. TMT provides multiplexing advantages and reduced technical variation but introduces specific challenges related to labeling efficiency, ratio compression, and batch effects.

Core Principles of TMT Labeling Chemistry

TMT reagents consist of three functional components: a mass reporter group, a mass normalizer group, and an amine-reactive group. The amine-reactive group covalently binds to primary amines on peptides, typically the N-terminus and lysine side chains. The reporter and normalizer groups are designed so that each tag variant has the same total mass, which means labeled peptides from different channels co-elute and appear as a single precursor peak in the mass spectrometer.

During MS/MS fragmentation, the reporter group is cleaved from the peptide, producing diagnostic ions at distinct masses that correspond to each TMT channel. The relative abundance of these reporter ions reflects the relative quantity of the peptide in each original sample.

Several factors critically influence labeling success. Sample pH is a major determinant of labeling efficiency. Peptide samples that remain acidic after resuspension in 50 mM HEPES buffer at pH 8.5 show low labeling efficiency and reduced median reporter ion intensities [12]. Increasing the buffer concentration to 500 mM HEPES for labeling consistently improves labeling efficiency and reduces missing data [12].

Protein-level TMT labeling presents additional challenges. Intact proteins tend to precipitate under labeling conditions, and side products from incomplete labeling or labeling of unintended residues increase sample complexity [5]. A benchmarking protocol for protein-level TMT labeling in top-down proteomics addresses these issues by removing large proteoforms before labeling, enabling characterization of small proteoforms below 35 kDa [5].

Experimental Design Considerations

Multiplex Layout and Channel Assignment

The number of available TMT channels determines how many samples can be analyzed in a single multiplex set. Common formats include 6-plex, 10-plex, 11-plex, 16-plex, and newer 35-plex reagents. The choice of format affects experimental design, cost, and statistical power.

When assigning samples to TMT channels, consider the following principles:

• Distribute biological replicates across channels instead of assigning all replicates of one condition to adjacent channels
• Balance conditions across batches when sample numbers exceed the multiplex capacity
• Include a common reference sample in each batch to enable cross-batch normalization
• Reserve channels for quality control samples if needed

The 16-plex format has been used extensively in large-scale studies. The Vanderbilt Memory and Aging Project used 23 16-plex TMT batches to analyze plasma samples from 336 participants, yielding 3764 unique protein identifications [10]. After rigorous quality control, 686 proteins were used as predictors in subsequent analyses [10].

Batch Effects and Study Design

Batch effects arise from systematic technical variation introduced during sample processing, labeling, or instrument analysis. These effects can confound biological comparisons if not properly controlled.

Strategies to manage batch effects include:

• Randomizing sample order across batches
• Including bridging samples or reference pools in every batch
• Using a common internal standard for normalization
• Recording batch information for statistical adjustment

The composition of TMT multiplexes influences quantification accuracy. Correlation of measured TMT reporter ratios with spiked-in standard peptide amounts is significantly lower for multiplexes composed of individual cerebrospinal fluid samples compared to those composed of aliquots of a single pool [16]. This finding demonstrates that heterogeneous sample composition affects TMT quantitation and must be considered in study design [16].

Sample Number and Input Requirements

The amount of protein required for TMT labeling depends on the sample type and experimental goals. Plant proteomics protocols require 0.1 mg protein per sample for reliable quantification of more than 8000 proteins [11]. For limited samples, optimized and scaled-down TMT labeling strategies have been developed to accommodate small input amounts [21].

For single-cell applications, the SCoPE2 protocol uses minimal proteomic sample preparation and isobaric mass tags for multiplexed analysis, enabling cost-effective quantification that can be automated and scaled to thousands of single cells [6]. The workflow allows analyzing approximately 200 single cells per 24 hours using standard commercial equipment [6].

Sample Preparation Workflow

Protein Extraction and Quantification

The first step in TMT proteomics is extracting proteins from the biological sample of interest. The extraction method must be compatible with downstream digestion and labeling steps. Common approaches include SDS-based extraction, urea-based lysis, and filter-assisted sample preparation.

For plant tissues, an optimized protocol uses SDS protein extraction followed by protein precipitation, digestion, TMT labeling, phosphopeptide enrichment, high pH reversed-phase fractionation, and LC-MS/MS analysis [11]. This protocol enables quantification of both phosphopeptides and non-phosphopeptides from the same samples [11].

For cerebrospinal fluid, comparative evaluation of filter-assisted sample preparation and in-solution protocols has been performed to address analytical challenges specific to this biofluid [16]. The choice of preparation method affects protein recovery, digestion efficiency, and downstream labeling performance.

Protein Digestion

Proteins must be digested into peptides before TMT labeling. Trypsin is the most commonly used protease, cleaving C-terminal to arginine and lysine residues. Digestion conditions, including enzyme-to-protein ratio, incubation time, and temperature, must be optimized for complete digestion.

Incomplete digestion produces missed cleavages that can affect quantification accuracy. Peptides with missed cleavage sites may label differently or produce altered reporter ion ratios. Quality control checks should include assessment of missed cleavage rates in the final dataset.

Peptide Cleanup and Resuspension

After digestion, peptides must be cleaned to remove salts, detergents, and other contaminants that interfere with labeling or mass spectrometry analysis. Solid-phase extraction or similar cleanup methods are typically used.

The resuspension buffer for TMT labeling must maintain proper pH. Peptide samples that remain acidic after resuspension in 50 mM HEPES buffer at pH 8.5 show low labeling efficiency [12]. Using 500 mM HEPES buffer for resuspension before labeling results in consistently higher labeling efficiency and lower missing data [12].

TMT Labeling Protocol

Reagent Preparation and Handling

TMT reagents are supplied as NHS ester derivatives that react with primary amines. The reagents must be dissolved in an appropriate solvent, typically anhydrous acetonitrile, immediately before use. Each channel of TMT reagent should be reconstituted to the same concentration to ensure equal labeling across samples.

The amount of TMT reagent needed depends on the peptide amount. Optimized protocols require lower tag-label concentrations while achieving consistently reliable data [17]. The protocol described in the Journal of Visualized Experiments provides detailed steps for extraction, quantification, precipitation, digestion, labeling, and data analysis of protein samples [17].

Labeling Reaction Conditions

The labeling reaction requires alkaline conditions, typically pH 8.0 to 8.5. Peptides are resuspended in HEPES buffer at the appropriate concentration, and TMT reagents are added. The reaction proceeds for a specified time, typically 1 to 2 hours at room temperature.

Sample pH is a critical factor in labeling efficiency [12]. If peptide samples remain acidic after resuspension, labeling efficiency decreases and reporter ion intensities are reduced [12]. Increasing buffer concentration in poorly labeled samples before relabeling can successfully rescue TMT labeling reactions [12].

Quenching and Sample Pooling

After the labeling reaction, excess TMT reagent must be quenched to prevent cross-labeling between samples. Hydroxylamine is commonly used for this purpose. Following quenching, samples from different channels are combined into a single multiplex sample.

Before combining full samples, a quality control step using a small aliquot from each sample within a TMT set is recommended [12]. This "Mixing QC" approach enables detection of TMT labeling issues by LC-MS/MS before combining the full samples, allowing salvage of poor TMT labeling reactions [12].

Quality Control Assessment

Several metrics should be assessed to evaluate labeling quality:

• Labeling efficiency, defined as the percentage of peptides with complete TMT modification
• Reporter ion missingness across channels
• Ratio of mean TMT set median reporter ion intensity to individual channel median reporter ion intensity
• Distribution of log2 reporter ion ratios in Mixing QC samples

These metrics provide multiple methods for assessing labeling quality before combining samples [12]. Implementing multiple assessment methods ensures robust TMT labeling for large-scale quantitative studies [12].

Fractionation and Mass Spectrometry Analysis

Peptide Fractionation

After pooling labeled samples, peptide fractionation reduces sample complexity and increases proteome depth. High pH reversed-phase fractionation is commonly used, either offline or online. The number of fractions collected depends on the experimental goals and instrument time available.

For phosphoproteomics applications, phosphopeptide enrichment is performed before fractionation or between fractionation steps [9]. The modular pipeline for TMT-DDA and LFQ-DIA integrates steps for scalable phosphoproteome profiling, including protein lysate extraction, cleanup, digestion, phosphopeptide enrichment, and TMT labeling [9].

Acquisition Strategies

Two primary acquisition strategies are used for TMT quantification: MS2 and MS3. In MS2 acquisition, reporter ions are measured from the fragmentation of the precursor peptide. In MS3 acquisition, an additional isolation and fragmentation step is performed to improve quantification accuracy.

Comparison of MS2 and MS3 acquisition strategies for cerebrospinal fluid TMT proteomics showed that the slope of measured versus expected ratios was improved by acquiring data in MS3 mode [16]. However, MS3 acquisition requires more instrument time and may reduce the number of peptides identified.

The choice between MS2 and MS3 depends on the balance between quantification accuracy and throughput. For studies where accurate fold-change estimation is critical, MS3 may be preferred despite the throughput cost.

Data-Dependent and Data-Independent Acquisition

TMT-based data-dependent acquisition (TMT-DDA) and data-independent acquisition-based label-free quantification (LFQ-DIA) have become leading workflows for deep proteome and phosphoproteome profiling [9]. Each approach has distinct advantages and limitations.

TMT-DDA selects precursor ions for fragmentation based on intensity, providing high-quality MS2 spectra for peptide identification. LFQ-DIA systematically fragments all precursor ions within a mass range, providing more complete coverage but requiring more sophisticated data analysis.

The choice between TMT-DDA and LFQ-DIA depends on sample number, input amount, and experimental goals [9]. Guidance for selecting the best workflow based on these factors is available from protocol descriptions [9].

Data Analysis Pipeline

Database Searching and Peptide Identification

The first step in data analysis is identifying peptides from the acquired mass spectra. Database search engines match experimental spectra against theoretical spectra generated from protein sequence databases. Several search engines support TMT data analysis, including FragPipe, MaxQuant, Proteome Discoverer, and pFind.

The pFind workflow, consisting of pParse for data preprocessing, pFind for database searching, and pQuant for quantitation, offers exceptional sensitivity, precision, and speed [20]. This streamlined workflow can be implemented in any laboratory for identifying peptides, proteins, or post-translational modifications, or for quantitation based on various labeling strategies including TMT and iTRAQ [20].

Quantification Tools and Platforms

Several computational platforms support TMT quantification. FragPipe coupled with TMT-Integrator provides integrative reports at the gene, protein, peptide, and post-translational modification site levels [14]. Benchmarking against MaxQuant and Proteome Discoverer showed that FragPipe coupled with TMT-Integrator quantifies more proteins in whole proteome datasets, quantifies more phosphorylated sites in phosphoproteome datasets, and delivers more robust quantification performance [14].

TMT-Integrator is integrated into the widely used FragPipe computational platform and is a core component of TMT and iTRAQ data analysis workflows [14]. The tool has been demonstrated using publicly available TMT datasets, including clear cell renal cell carcinoma whole proteome and phosphoproteome datasets from the Clinical Proteomic Tumor Analysis Consortium [14].

Normalization Strategies

Normalization corrects for systematic biases introduced during sample processing and analysis. Several normalization approaches are available for TMT data:

• Median normalization adjusts channel intensities to have the same median
• Quantile normalization forces identical distributions across channels
• Total peptide amount normalization scales channels based on total signal
• TAMPOR normalization uses a reference-based approach

For cerebrospinal fluid TMT proteomics, correlation of measured reporter ratios with expected values was improved by applying median and quantile normalization [16]. The choice of normalization method affects downstream statistical analysis and biological interpretation.

Statistical Analysis and Interpretation

After normalization, statistical analysis identifies proteins with significant abundance changes between conditions. Common approaches include linear mixed-effects models, t-tests with multiple testing correction, and analysis of variance.

False discovery rate correction is essential when testing thousands of proteins simultaneously. The Vanderbilt Memory and Aging Project applied false discovery rate correction to identify plasma proteins associated with cognitive decline [10]. Linear mixed-effects regressions related protein levels to neuropsychological outcomes in fully adjusted models [10].

For studies examining post-translational modifications, site-level quantification requires additional considerations. Phosphorylation site localization confidence and site-level normalization must be addressed in the analysis pipeline.

Quality Control and Reproducibility

Mixing QC Samples

The Mixing QC approach embeds a quality control sample containing a small aliquot from each sample within a TMT set [12]. This QC sample is analyzed by LC-MS/MS before combining the full samples, enabling detection of labeling issues early in the workflow [12].

Metrics assessed in Mixing QC samples include labeling efficiency, reporter ion missingness, the ratio of mean TMT set median reporter ion intensity to individual channel median reporter ion intensity, and the distribution of log2 reporter ion ratios [12]. These metrics provide multiple methods for assessing labeling quality [12].

Labeling Efficiency Assessment

Labeling efficiency is a critical quality metric for TMT experiments. Low labeling efficiency produces peptides with missed modifications, which complicates quantification and reduces data completeness.

Sample pH is a critical factor in labeling efficiency [12]. Peptide samples that remain acidic after resuspension in 50 mM HEPES buffer at pH 8.5 show low labeling efficiency and relatively low median reporter ion intensities [12]. Resuspending peptides in 500 mM HEPES buffer for TMT labeling results in consistently higher labeling efficiency and lower missing data [12].

Reproducibility Measures

Reproducibility in TMT experiments depends on consistent sample processing, labeling, and analysis. Technical replicates, where the same sample is processed multiple times, assess the variability introduced by the workflow. Biological replicates assess the variability inherent in the biological system.

For large-scale studies spanning multiple TMT batches, cross-batch reproducibility is essential. The Vanderbilt Memory and Aging Project used 23 16-plex TMT batches and applied rigorous quality control to yield 686 proteins used as predictors in subsequent analyses [10]. This quality control process ensured that only reliably quantified proteins were included in downstream statistical modeling [10].

Common Failure Patterns and Troubleshooting

Low Labeling Efficiency

Low labeling efficiency is a common problem in TMT experiments. The primary cause is inadequate sample pH during the labeling reaction. Peptide samples that remain acidic after resuspension in 50 mM HEPES buffer at pH 8.5 show low labeling efficiency [12].

Troubleshooting steps include:

• Verify peptide resuspension pH before adding TMT reagents
• Increase buffer concentration to 500 mM HEPES for labeling
• Test labeling efficiency with a Mixing QC sample before combining full samples
• Consider relabeling with higher buffer concentration if initial labeling fails

Relabeling does not necessarily rescue TMT reactions [12]. Increasing the buffer concentration in poorly labeled samples before relabeling resulted in successful rescue of TMT labeling reactions [12].

Ratio Compression

Ratio compression occurs when reporter ion ratios are compressed toward one, reducing the apparent magnitude of fold changes. This effect is caused by co-isolation of contaminating peptides during precursor selection.

MS3 acquisition reduces ratio compression by providing an additional level of isolation. Comparison of MS2 and MS3 acquisition strategies showed that the slope of measured versus expected ratios was improved by acquiring data in MS3 mode [16].

Missing Reporter Ions

Missing reporter ions occur when a peptide is identified but reporter ion intensities are absent or below detection limits for some channels. This problem reduces data completeness and complicates statistical analysis.

Factors contributing to missing reporter ions include low labeling efficiency, low peptide abundance, and ion suppression. Using 500 mM HEPES buffer for labeling reduces missing data [12].

Batch Effects

Batch effects introduce systematic variation that can confound biological comparisons. The heterogeneous composition of individual samples within a TMT multiplex influences quantitation accuracy [16].

Strategies to address batch effects include:

• Including common reference samples in each batch
• Randomizing sample assignment across batches
• Applying appropriate normalization methods
• Including batch as a covariate in statistical models

Limitations and Interpretation Boundaries

Quantification Accuracy

TMT quantification provides relative instead of absolute protein abundances. The measured reporter ion ratios reflect relative quantities between samples within a multiplex set. Absolute quantification requires spiking in known amounts of standard peptides or proteins.

The accuracy of TMT quantification depends on sample composition. Correlation of measured TMT reporter ratios with spiked-in standard peptide amounts was significantly lower for TMT multiplexes composed of individual cerebrospinal fluid samples compared with those composed of aliquots of a single pool [16]. This finding demonstrates that heterogeneous sample composition influences TMT quantitation [16].

Protein-Level Labeling Challenges

Protein-level TMT labeling for top-down proteomics faces specific challenges. Intact proteins tend to precipitate under labeling conditions, and side products from incomplete labeling or labeling of unintended residues increase sample complexity [5].

A benchmarking protocol for protein-level TMT labeling addresses these challenges by removing large proteoforms before labeling, enabling characterization of small proteoforms below 35 kDa [5]. The protocol provides guidelines for isobaric chemical tag quantification in top-down proteomics [5].

Dynamic Range Limitations

Mass spectrometry-based proteomics has limited dynamic range compared to the range of protein abundances in biological samples. Low-abundance proteins may not be detected, particularly in complex samples.

The use of isobaric carriers can enhance peptide sequence identification in single-cell applications [6]. The SCoPE2 protocol uses an isobaric carrier to enhance peptide sequence identification, enabling quantification of over 1000 proteins per cell [6].

Data Sharing and Reproducibility

Reproducibility in proteomics requires transparent reporting of experimental details and data availability. Public data repositories support data sharing and reanalysis.

The FAIR Guiding Principles provide a framework for making data findable, accessible, interoperable, and reusable [4]. These principles apply to proteomics data, including raw mass spectrometry files, processed data, and metadata.

Data sharing policies may apply to research funded by specific agencies. The NIH Genomic Data Sharing Policy outlines expectations for data sharing in genomic research [3]. Researchers should be aware of applicable policies and ensure compliance.

Professional Escalation Criteria

Certain situations warrant consultation with specialized expertise or escalation to facility staff. The following criteria indicate when professional assistance is appropriate:

• Persistent low labeling efficiency despite buffer optimization
• Unexpected patterns in quality control metrics
• Instrument performance issues affecting data quality
• Statistical challenges requiring specialized bioinformatics support
• Regulatory or compliance questions regarding data sharing

For complex experimental designs, consultation with a proteomics facility or bioinformatics core can prevent costly errors. Facilities often provide guidance on experimental design, sample preparation, and data analysis.

Records and Documentation

Maintaining detailed records is essential for reproducible TMT proteomics. The following information should be documented for each experiment:

• Sample source, collection date, and storage conditions
• Protein extraction and digestion protocols
• TMT reagent lot numbers and channel assignments
• Labeling conditions, including buffer concentration and pH
• Quality control metrics, including labeling efficiency and reporter ion intensities
• Fractionation and acquisition parameters
• Data analysis software versions and parameters

This documentation supports troubleshooting, data interpretation, and publication requirements.

Frequently Asked Questions

What is the difference between TMT and iTRAQ labeling?

TMT and iTRAQ are both isobaric labeling strategies for multiplexed quantitative proteomics. TMT reagents are available in formats up to 16-plex and newer 35-plex versions, while iTRAQ is available in 4-plex and 8-plex formats. Both approaches use isobaric tags that produce distinct reporter ions upon fragmentation. The choice between them depends on multiplexing needs, instrument compatibility, and reagent availability. TMT-Integrator processes quantitation results from both TMT and iTRAQ experiments [14].

How much protein is needed for TMT labeling?

The required protein amount depends on the sample type and experimental goals. Plant proteomics protocols require 0.1 mg protein per sample for reliable quantification of more than 8000 proteins [11]. For limited samples, optimized and scaled-down TMT labeling strategies have been developed to accommodate small input amounts [21]. Single-cell applications use specialized protocols that enable analysis of individual cells [6].

What causes low TMT labeling efficiency?

Sample pH is a critical factor in labeling efficiency [12]. Peptide samples that remain acidic after resuspension in 50 mM HEPES buffer at pH 8.5 show low labeling efficiency and relatively low median reporter ion intensities [12]. Increasing buffer concentration to 500 mM HEPES for labeling results in consistently higher labeling efficiency and lower missing data [12].

How do I choose between MS2 and MS3 acquisition for TMT quantification?

MS3 acquisition improves quantification accuracy by reducing ratio compression but requires more instrument time. Comparison of MS2 and MS3 acquisition strategies for cerebrospinal fluid TMT proteomics showed that the slope of measured versus expected ratios was improved by acquiring data in MS3 mode [16]. The choice depends on the balance between quantification accuracy and throughput.

What normalization method should I use for TMT data?

Several normalization approaches are available, including median, quantile, total peptide amount, and TAMPOR normalization. For cerebrospinal fluid TMT proteomics, correlation of measured reporter ratios with expected values was improved by applying median and quantile normalization [16]. The choice of normalization method depends on the data characteristics and experimental design.

How do I handle batch effects in large-scale TMT studies?

Batch effects can be managed by including common reference samples in each batch, randomizing sample assignment across batches, applying appropriate normalization methods, and including batch as a covariate in statistical models. Large-scale studies such as the Vanderbilt Memory and Aging Project used 23 16-plex TMT batches and applied rigorous quality control to ensure reliable quantification [10].

What is the Mixing QC approach?

The Mixing QC approach embeds a quality control sample containing a small aliquot from each sample within a TMT set [12]. This QC sample is analyzed by LC-MS/MS before combining the full samples, enabling detection of TMT labeling issues early in the workflow [12]. Metrics assessed include labeling efficiency, reporter ion missingness, and the distribution of log2 reporter ion ratios [12].

Can TMT be used for single-cell proteomics?

Yes, TMT can be used for single-cell proteomics. The SCoPE2 protocol uses an isobaric carrier to enhance peptide sequence identification, enabling quantification of over 1000 proteins per cell [6]. The workflow allows analyzing approximately 200 single cells per 24 hours using standard commercial equipment [6].

Related Bioinformatics Guides

• Alternative Splicing Analysis from RNA-Seq Data
• Codon Usage Bias Analysis for Recombinant Vaccine Design
• Single-Cell RNA-Seq Analysis Pipelines for Veterinary Immunology
• Structural and Evolutionary Analysis of Rabies Virus Glycoprotein: Implications for Vaccine Design
• Structural and Computational Analysis of African Swine Fever Virus Capsid Proteins for Antiviral Drug Design

References and Further Reading

• EMBL-EBI Training. European Bioinformatics Institute.
• NCBI Data Resources. National Center for Biotechnology Information.
• Genomic Data Sharing Policy. National Institutes of Health.
• The FAIR Guiding Principles. Scientific Data.
• A benchmarking protocol for intact protein-level Tandem Mass Tag (TMT) labeling for quantitative top-down proteomics.. MethodsX, 2022.
• Multiplexed single-cell proteomics using SCoPE2.. Nature protocols, 2021.
• Quantitative Translation Proteomics Using mePROD.. Methods in molecular biology (Clifton, N.J.), 2022.
• Studying Autophagy Using a TMT-Based Quantitative Proteomics Approach.. Methods in molecular biology (Clifton, N.J.), 2022.
• Protocol for high-throughput semi-automated label-free- or TMT-based phosphoproteome profiling.. STAR protocols, 2023.
• LC-MS/MS proteomics identifies plasma proteins related to cognition over 9-year follow-up.. Alzheimer's & dementia : the journal of the Alzheimer's Association, 2025.
• Tandem Mass Tag-Based Phosphoproteomics in Plants.. Methods in molecular biology (Clifton, N.J.), 2023.
• Assessment of TMT Labeling Efficiency in Large-Scale Quantitative Proteomics: The Critical Effect of Sample pH.. ACS omega, 2021.
• Label-Free Target Discovery Strategy for Natural Active Products.. 2026.
• Analysis of isobaric quantitative proteomic data using TMT-Integrator and FragPipe computational platform.. 2026.
• In Depth Characterization of the Promoter Proximal Proteome of Single Copy Locus FOXP2.. 2026.
• Optimized sample preparation and data analysis for TMT proteomic analysis of cerebrospinal fluid applied to the identification of Alzheimer’s disease biomarkers. Clinical Proteomics, 2022.
• TMT Sample Preparation for Proteomics Facility Submission and Subsequent Data Analysis. Journal of Visualized Experiments, 2020.
• TMT Diversity and the Financial Performance of Listed Chinese Companies: Three-Way Interaction Analysis of Innovativeness and Government R&D Subsidies. Syst., 2025.
• Genome-Wide Identification and an Evolution Analysis of Tonoplast Monosaccharide Transporter (TMT) Genes in Seven Gramineae Crops and Their Expression Profiling in Rice. Genes, 2023.
• How to use open-pFind in deep proteomics data analysis?- A protocol for rigorous identification and quantitation of peptides and proteins from mass spectrometry data. Biophysics Reports, 2021.
• Opt-TMT: An optimized and scaled-down TMT labeling strategy for limited sample. Journal of Proteomics, 2026.

This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.