Bottom-Up Proteomics: Principles, Workflow, and Applications

Bottom-up proteomics, also called shotgun proteomics, is the prevailing mass spectrometry strategy in which proteins are first hydrolyzed into peptides and then analyzed to infer the identity, quantity, and modification state of the original proteins. This approach dominates contemporary proteomics because peptides are more amenable to liquid chromatography separation and tandem mass spectrometry fragmentation than intact proteins. For researchers, analysts, and life-science professionals, understanding the bottom-up workflow is essential for designing experiments that yield reproducible, interpretable, and biologically meaningful results. This article explains the principles of bottom-up proteomics, walks through each stage of the workflow, compares it with top-down proteomics, and provides practical guidance on quality control, troubleshooting, and data interpretation.

The Core Principle of Bottom-Up Proteomics

Bottom-up proteomics rests on a simple but powerful logic: instead of analyzing intact proteins, the researcher digests proteins into peptides, separates those peptides by liquid chromatography, and then uses tandem mass spectrometry to generate fragmentation spectra that can be matched to peptide sequences. Because peptides are smaller and more uniform in their physicochemical properties than proteins, they separate more predictably on reversed-phase columns and fragment more reproducibly in the mass spectrometer. This makes the bottom-up strategy the workhorse for large-scale protein identification and quantification across cells, tissues, and body fluids.

The term "bottom-up" refers to the analytical direction: the experiment starts at the peptide level, the bottom of the protein hierarchy, and reconstructs protein-level information through computational inference. This contrasts with top-down proteomics, which introduces intact proteins into the mass spectrometer and fragments them directly. The choice between these strategies has profound consequences for sample preparation, instrumentation requirements, data analysis, and the types of biological questions that can be answered.

Proteomics as a field studies protein structure and function at large scale through protein identification and quantification. The applications extend from simple protein cataloging to the characterization of proteoforms, protein-protein interactions, structural alterations, absolute and relative quantification, post-translational modifications, and protein stability. Each of these applications places different demands on the workflow, and the bottom-up strategy offers the flexibility to address most of them with standard laboratory equipment and widely available software.

At a Glance: Bottom-Up Versus Top-Down Proteomics

The table below summarizes the key differences between bottom-up and top-down proteomics across the dimensions that matter most for experimental design.

The choice between bottom-up and top-down is not always exclusive. Integrated approaches that combine both strategies can produce a more precise proteoform landscape than either method alone. For example, combining top-down and bottom-up proteomics has been shown to improve the characterization quality for approximately 35 percent of identified proteoforms containing mass shifts in studies of the protein corona formed on nanoparticles. This integrated strategy enables the discovery and precise characterization of potential proteoform biomarkers that would be missed by either approach in isolation.

The Bottom-Up Proteomics Workflow

A typical bottom-up proteomic workflow consists of three major stages: sample preparation, LC-MS/MS analysis, and data analysis. Each stage contains multiple decision points where the researcher must balance throughput, reproducibility, depth of coverage, and quantitative accuracy. The following sections describe each stage in detail, with attention to the practical choices that determine experimental success.

Sample Preparation: The Critical First Stage

Sample preparation is the stage that most directly affects the overall efficiency of a proteomic study. It is laborious, prone to errors, and historically has shown low reproducibility and throughput. While LC-MS/MS and data analysis techniques have been intensively developed, sample preparation remains the main challenge in many applications. The quality of the data produced by the mass spectrometer can never exceed the quality of the peptides introduced into it, so careful attention to this stage is not optional.

The first decision in sample preparation is the choice of lysis buffer and protein extraction method. The lysis buffer must solubilize proteins from the biological matrix while simultaneously inactivating endogenous proteases that would otherwise degrade the sample unpredictably. Detergents are commonly used to solubilize membrane proteins, but they interfere with downstream chromatography and ionization and must be removed before LC-MS/MS analysis. Chaotropes such as urea and thiourea denature proteins and improve digestion efficiency but can carbamylate peptides if used at elevated temperatures for extended periods. Reducing agents break disulfide bonds, and alkylating agents prevent their reformation, ensuring that cysteine-containing peptides are amenable to database searching.

Protein cleanup and digestion are the next critical steps. In-solution digestion and filter-aided sample preparation are the typical and widely used methods. In-solution digestion is straightforward: proteins are denatured, reduced, alkylated, and then incubated with a protease, most commonly trypsin, which cleaves C-terminal to arginine and lysine residues. The resulting peptides are then desalted and concentrated before LC-MS/MS analysis. This method is simple and scalable but requires careful removal of detergents and other interfering substances.

Filter-aided sample preparation, commonly abbreviated FASP, uses ultrafiltration units with membranes having large molecular mass cut-offs. The method allows buffer exchange, detergent removal, protein digestion, and peptide collection to occur on a single filter device. FASP is applicable to a variety of sample types and produces high-quality peptides. It permits digestion with a variety of enzymes, allows straightforward monitoring of protein-to-peptide conversion, and uniquely enables consecutive cleavage with several proteases and separation of peptide fractions. Successful application of FASP requires optimized properties of the sample lysate and its amount, the use of appropriate ultrafiltration units, and well-selected conditions for protein digestion.

In the past decade, novel methods have been developed to improve and facilitate the entire sample preparation process or to integrate sample preparation with fractionation. These include on-membrane digestion, bead-based digestion, immobilized enzymatic digestion, and suspension trapping. The goals of these methods are to reduce time, increase throughput, and improve reproducibility. Bead-based digestion, for example, uses magnetic or polymeric beads to capture proteins and enzymes, enabling rapid buffer exchange and digestion in a single tube. Suspension trapping combines protein precipitation, digestion, and peptide cleanup in a single device, reducing sample loss and hands-on time.

Automation is increasingly important for large-scale studies. Fully integrated, automated sample preparation platforms that cover the entire process from biological sample input to mass spectrometry-ready peptide output have been developed and applied to a multitude of biological samples. These platforms achieve high intra-plate and inter-plate reproducibility as well as longitudinal consistency, and they surpass established manual and semi-automated workflows while improving time efficiency. For clinical applications and drug development, where high throughput and quantitative accuracy are indispensable, automated sample preparation is becoming a practical necessity.

Protein Quantification Before Digestion

Accurate protein quantification before digestion is essential for loading consistent amounts of peptides onto the LC-MS/MS system. Total protein measurement methods include bicinchoninic acid assay, Bradford assay, and absorbance at 280 nanometers. Each method has biases: the bicinchoninic acid assay is compatible with detergents but sensitive to reducing agents, the Bradford assay is rapid but variable across protein types, and absorbance at 280 nanometers requires knowledge of the extinction coefficient and is confounded by nucleic acids. The choice of quantification method should be matched to the lysis buffer composition and the expected protein concentration range.

For label-free quantification workflows, equal protein amounts across samples are critical for accurate differential expression analysis. Loading unequal amounts introduces systematic bias that can be misinterpreted as biological variation. For tandem mass tag or isobaric labeling workflows, protein quantification before labeling ensures that each sample contributes equally to the multiplexed analysis, reducing ratio compression and improving quantitative accuracy.

Enzymatic Digestion Choices

Trypsin is the default protease for bottom-up proteomics because it produces peptides with C-terminal arginine or lysine residues, which ionize well in positive mode electrospray and fragment predictably in collision-induced dissociation. The average peptide length produced by trypsin digestion, approximately 7 to 15 amino acids, is well suited to reversed-phase chromatography and database searching.

Other proteases expand the sequence coverage and enable the analysis of post-translational modifications that fall within tryptic peptides poorly. Lys-C cleaves C-terminal to lysine residues and is often used in combination with trypsin to improve digestion efficiency, particularly for membrane proteins. Glu-C cleaves C-terminal to glutamic acid and aspartic acid under appropriate conditions, generating larger peptides that can improve sequence coverage in regions where tryptic peptides are too short or too long. Chymotrypsin cleaves C-terminal to aromatic residues and provides complementary coverage. The choice of protease or protease combination should be guided by the biological question, the expected protein classes, and the modification sites of interest.

Peptide Fractionation and Enrichment

Complex proteomes contain hundreds of thousands of peptides, far more than can be sequenced in a single LC-MS/MS run. Peptide fractionation before LC-MS/MS reduces complexity and increases the depth of proteome coverage. Common fractionation methods include strong cation exchange chromatography, high-pH reversed-phase chromatography, and isoelectric focusing. These methods separate peptides by charge or hydrophobicity, and each fraction is then analyzed separately by LC-MS/MS. The trade-off is clear: more fractions mean deeper coverage but also more instrument time and more data to process.

For post-translational modification analysis, enrichment is often required before LC-MS/MS. Phosphopeptide enrichment using immobilized metal affinity chromatography or titanium dioxide is standard for phosphoproteomics. Glycopeptide enrichment using lectin affinity or hydrazide chemistry enables glycoproteomics. Ubiquitin remnant peptide enrichment using anti-diGly antibodies allows the identification of ubiquitination sites. Each enrichment strategy introduces its own biases and requires careful optimization to balance specificity and sensitivity.

LC-MS/MS Analysis

Liquid chromatography-tandem mass spectrometry is the analytical engine of bottom-up proteomics. Peptides are separated on a reversed-phase column using a gradient of increasing organic solvent, eluted into the electrospray source, ionized, and introduced into the mass spectrometer. The mass spectrometer first measures the mass-to-charge ratio of intact peptide ions in a survey scan, then selects precursor ions for fragmentation, and finally measures the mass-to-charge ratios of the resulting fragment ions. The fragmentation spectra, or tandem mass spectra, are then matched to peptide sequences.

Nanoscale liquid chromatography coupled to tandem mass spectrometry, abbreviated nanoLC-MS/MS, is the standard configuration for bottom-up proteomics. The reduced column diameter, typically 75 micrometers, increases the concentration of eluting peptides and improves ionization efficiency, leading to higher sensitivity. The trade-off is lower flow rates and longer gradients, which increase analysis time but improve chromatographic resolution.

Data acquisition strategies fall into two broad categories: data-dependent acquisition and data-independent acquisition. In data-dependent acquisition, the mass spectrometer selects the most abundant precursor ions from each survey scan for fragmentation. This strategy is simple and widely used but suffers from stochastic sampling, where low-abundance peptides are missed in some runs and detected in others, reducing reproducibility across replicates. In data-independent acquisition, the mass spectrometer systematically fragments all precursor ions within defined isolation windows, regardless of abundance. This strategy provides more complete and reproducible coverage but generates highly complex fragment spectra that require specialized analysis software.

Data-independent acquisition has emerged as a powerful technology for high-throughput, accurate, and reproducible quantitative proteomics. Acquisition schemes are categorized based on the design of precursor isolation windows, including wide-window, overlapping-window, narrow-window, scanning quadrupole-based, and parallel accumulation-serial fragmentation-enhanced methods. For data analysis, major strategies include spectrum reconstruction, sequence-based search, library-based search, de novo sequencing, and sequencing-independent approaches. The generation and optimization of spectral libraries, which are critical resources for data-independent acquisition analysis, require careful attention to sample complexity, instrument settings, and library depth.

Collision Energy Optimization

The choice of collision energy in tandem mass spectrometric experiments has an outstanding role in the bottom-up approach. Collision energy determines the extent of peptide fragmentation and therefore the quality of the tandem mass spectra. Too little energy produces incomplete fragmentation with few informative ions, while too much energy produces excessive fragmentation with loss of sequence-informative ions. Modern instruments use collision energy ramps or calculated values based on precursor mass and charge, but optimization is still required for specific instrument platforms, peptide modifications, and protease choices.

Collision energy optimization strategies have been developed to fully exploit the potential of mass spectrometry-based proteomics techniques. These strategies include stepped collision energies, where multiple energies are applied in a single fragmentation event, and machine learning-based prediction of optimal energies based on peptide properties. Comparing results from different studies or different instruments requires careful attention to collision energy settings, and methodology transfer between laboratories can be facilitated by measuring reference species with known fragmentation behavior.

Data Analysis and Protein Inference

Data analysis is the computational stage where tandem mass spectra are converted into peptide identifications, peptide identifications are assembled into protein identifications, and protein quantities are estimated. This stage requires specialized software, reference databases, and careful statistical validation.

Database Searching

The most common approach to peptide identification is database searching. Each tandem mass spectrum is compared against theoretical spectra generated from an in silico digestion of a protein sequence database. The search engine scores the match between the observed and theoretical spectra, and the best match is reported with a statistical confidence score. Common search engines include Proteome Discoverer, MaxQuant, Mascot, and MS-GF+, each with its own scoring algorithm and parameter requirements.

The choice of protein sequence database is critical. For well-annotated organisms, the reference proteome from a major database such as the National Center for Biotechnology Information provides a complete and accurate set of protein sequences. For less-characterized organisms or for samples containing multiple organisms, the database must include all expected protein sequences, and the risk of false identifications increases with database size. For multi-organism systems, such as host-pathogen interactions, the database must include protein sequences from all relevant taxa, and the analysis must account for the possibility of homologous peptides shared between organisms.

False Discovery Rate Control

Peptide and protein identifications must be validated statistically to control the false discovery rate. The standard approach uses a target-decoy database search strategy, where the search is performed against both the target protein sequences and a set of reversed or shuffled decoy sequences. The number of decoy identifications at a given score threshold estimates the number of false target identifications, allowing the calculation of a false discovery rate. A false discovery rate of 1 percent at the peptide level and 1 percent at the protein level is a common standard for discovery proteomics.

The false discovery rate must be interpreted in the context of the experiment. For hypothesis-generating discovery studies, a 1 percent false discovery rate is appropriate. For targeted validation studies, a more stringent threshold may be required. For clinical applications, where individual identifications may influence patient management decisions, the false discovery rate should be as low as practically achievable, and orthogonal validation by an independent method is strongly recommended.

Protein Inference and Quantification

The assignment of peptides to proteins is complicated by the existence of shared peptides, which are peptides that match multiple proteins or protein isoforms. The parsimony principle is used to assemble the minimal set of proteins that explains all observed peptides. Proteins that are inferred solely from shared peptides are reported with lower confidence than proteins with unique peptides. The distinction between protein groups, which contain proteins that cannot be distinguished based on the observed peptides, and individual proteins is an important concept in interpreting proteomics results.

Quantification strategies fall into two broad categories: label-free and labeled. Label-free quantification compares peptide peak intensities or spectral counts across runs. It is simple, cost-effective, and applicable to any sample type, but it requires careful normalization and is sensitive to run-to-run variation. Labeled quantification uses chemical or metabolic tags to distinguish samples within a single run. Tandem mass tag and isobaric tags for relative and absolute quantification enable multiplexed analysis of up to 16 or more samples, improving throughput and reducing missing values. Metabolic labeling with stable isotope labeling by amino acids in cell culture enables accurate relative quantification in cell culture systems but is not applicable to human samples or most animal tissues.

Spectral Libraries and Data-Independent Acquisition Analysis

Data-independent acquisition data analysis often relies on spectral libraries, which are collections of peptide fragmentation spectra with associated retention times and quantitative information. Spectral libraries can be generated from data-dependent acquisition runs of the same sample type or from public repositories. The quality and completeness of the spectral library directly affect the sensitivity and accuracy of data-independent acquisition analysis. Library-free approaches, which use sequence-based search or de novo sequencing, are also available and are particularly useful when no suitable library exists.

Publicly available benchmark datasets covering global proteomics and phosphoproteomics facilitate the performance evaluation of various software tools and analysis workflows. These datasets provide standardized inputs and expected outputs, allowing researchers to compare the performance of different analysis strategies on their own data.

Practical Implementation Steps

The following steps provide a practical framework for implementing a bottom-up proteomics experiment. These steps are applicable to a wide range of sample types and biological questions.

Step 1: Define the Biological Question and Experimental Design

The biological question determines every downstream choice. For global protein profiling, a label-free or tandem mass tag workflow with deep fractionation is appropriate. For targeted quantification of specific proteins, a selected reaction monitoring or parallel reaction monitoring workflow is more suitable. For post-translational modification analysis, enrichment steps must be included. The experimental design must include appropriate biological replicates, technical replicates, and controls to support the intended statistical analysis.

Step 2: Select the Sample Preparation Method

Choose the sample preparation method based on sample type, throughput requirements, and available equipment. For small numbers of samples, in-solution digestion or filter-aided sample preparation is appropriate. For large cohorts, automated platforms or suspension trapping may be necessary. The chosen method must be validated for the specific sample type, and the protein quantification method must be compatible with the lysis buffer.

Step 3: Optimize Digestion Conditions

Optimize the enzyme-to-protein ratio, digestion time, and digestion temperature for the specific sample type. Verify digestion completeness by monitoring protein-to-peptide conversion, for example by SDS-PAGE or by measuring the peptide-to-protein ratio. For filter-aided sample preparation, confirm that the ultrafiltration membrane retains proteins during buffer exchange and releases peptides efficiently during the collection step.

Step 4: Perform Quality Control Checks

Before LC-MS/MS analysis, assess peptide quantity and quality. Measure peptide concentration using a compatible assay, and check a small aliquot by LC-MS/MS to evaluate chromatography, ionization, and fragmentation. Verify that the base peak chromatogram shows a reasonable distribution of peptide peaks and that the number of identified peptides is consistent with expectations for the sample type and instrument.

Step 5: Acquire LC-MS/MS Data

Set up the LC-MS/MS method with appropriate gradient length, collision energy settings, and data acquisition mode. For discovery experiments, data-dependent acquisition with a 1 percent false discovery rate target is standard. For quantitative experiments requiring high reproducibility, data-independent acquisition may be preferable. Include quality control samples at regular intervals throughout the acquisition to monitor instrument performance.

Step 6: Analyze Data and Validate Results

Process the raw data with the chosen analysis software, using appropriate search parameters and false discovery rate thresholds. Inspect the distribution of identification scores, the number of peptides per protein, and the reproducibility across replicates. Validate key findings by an orthogonal method, such as western blotting, enzyme-linked immunosorbent assay, or targeted mass spectrometry, particularly for clinical or high-impact applications.

Step 7: Interpret and Report Results

Interpret the results in the biological context, using pathway analysis and gene ontology tools to identify enriched functions and processes. Report the experimental details, including sample preparation method, digestion conditions, LC-MS/MS parameters, data analysis software, and false discovery rate thresholds, to enable reproducibility. Deposit the raw data and analysis files in a public repository to comply with data sharing expectations.

Records and Measurements

Systematic record keeping is essential for reproducible bottom-up proteomics. The following records should be maintained for each experiment.

Sample Metadata

Record the sample source, collection date, storage conditions, and any treatments or perturbations. For animal or human samples, record relevant demographic and clinical information. For multi-organism systems, record the identity and state of each organism. This metadata is essential for interpreting the results and for complying with data sharing policies.

Sample Preparation Records

Record the lysis buffer composition, protein quantification method and results, digestion enzyme and conditions, cleanup method, and fractionation or enrichment details. Record the final peptide concentration and the volume loaded onto the LC-MS/MS system. These records enable troubleshooting and facilitate method transfer between laboratories.

Instrument and Acquisition Records

Record the LC-MS/MS instrument, column type and batch, gradient profile, collision energy settings, and data acquisition mode. Record the quality control results, including the base peak chromatogram, total ion current, and the number of identified peptides or proteins for each run. Monitor these metrics over time to detect instrument drift or column degradation.

Data Analysis Records

Record the software version, search parameters, protein sequence database and version, false discovery rate thresholds, and quantification method. Record the number of identified peptides and proteins, the number of quantified proteins, and the missing value rates. These records enable the analysis to be reproduced and the results to be compared across studies.

Common Failure Patterns and Troubleshooting

Several failure patterns recur in bottom-up proteomics. Recognizing these patterns and understanding their causes is essential for efficient troubleshooting.

Low Peptide Yield

Low peptide yield can result from incomplete protein extraction, inefficient digestion, or sample loss during cleanup. Verify that the lysis buffer is appropriate for the sample type and that the protein quantification method is accurate. Check digestion completeness by monitoring protein-to-peptide conversion. Minimize sample transfers and use low-binding tubes to reduce loss.

Poor Chromatography

Poor chromatography, characterized by broad peaks, tailing, or shifting retention times, can result from column degradation, sample contamination, or inappropriate gradient conditions. Replace the column if performance does not improve after cleaning. Verify that the sample is free of detergents and other contaminants that interfere with reversed-phase chromatography.

Low Identification Rates

Low identification rates can result from suboptimal collision energy, insufficient chromatographic resolution, or an incomplete protein sequence database. Optimize collision energy settings for the specific instrument and sample type. Increase the gradient length or add fractionation to reduce sample complexity. Verify that the database includes all expected protein sequences.

Poor Quantitative Reproducibility

Poor quantitative reproducibility can result from unequal sample loading, run-to-run instrument variation, or incomplete digestion. Verify that equal protein amounts are loaded for label-free workflows. Use quality control samples to monitor instrument performance. Optimize digestion conditions to ensure complete and consistent protein-to-peptide conversion.

Contamination and Carryover

Contamination from keratins and other environmental proteins is a common problem in bottom-up proteomics. Minimize sample handling, use clean gloves and lab coats, and avoid opening tubes unnecessarily. Carryover between runs can be reduced by including blank runs and wash steps in the acquisition method.

Limitations of Bottom-Up Proteomics

Bottom-up proteomics has inherent limitations that must be considered when interpreting results.

Loss of Proteoform Information

Because proteins are digested into peptides before analysis, the connection between a specific peptide and its parent proteoform is often lost. Peptides shared between multiple proteoforms cannot be unambiguously assigned, and the combination of post-translational modifications on a single protein molecule cannot be determined from peptide-level data. This limitation is particularly important for studying proteoforms, where the precise combination of modifications determines function.

Sequence Coverage Limitations

Bottom-up proteomics rarely achieves complete sequence coverage of any protein. Transmembrane domains, highly hydrophobic regions, and regions lacking protease cleavage sites are often underrepresented. Post-translational modifications in these regions may be missed, and the absence of a peptide cannot be interpreted as evidence that a modification is absent.

Dynamic Range Limitations

The dynamic range of protein abundances in biological samples spans many orders of magnitude, far exceeding the dynamic range of the mass spectrometer. High-abundance proteins suppress the detection of low-abundance proteins, and abundant peptides dominate data-dependent acquisition selection. Depletion of high-abundance proteins or extensive fractionation can partially address this limitation but introduces its own biases.

Quantification Accuracy

Label-free quantification is sensitive to run-to-run variation and missing values. Isobaric labeling reduces missing values but can suffer from ratio compression due to co-isolation of contaminating ions. Absolute quantification requires stable isotope-labeled standards for each target peptide, which is impractical for large-scale studies.

Welfare and Safety Context

Bottom-up proteomics involves the use of chemicals, biological samples, and laboratory equipment that require appropriate safety precautions.

Chemical Safety

Lysis buffers often contain chaotropes, detergents, reducing agents, and protease inhibitors that can be hazardous. Urea should be handled with care, as it can generate cyanate, which carbamylates proteins. Acrylamide, used in some gel-based workflows, is a neurotoxin. Organic solvents used in chromatography, such as acetonitrile and methanol, are flammable and should be used in ventilated areas. All chemicals should be handled according to the safety data sheet and institutional guidelines.

Biological Safety

Biological samples may contain infectious agents, particularly when working with blood, tissue, or multi-organism systems. Samples should be handled in appropriate biosafety cabinets, and all waste should be decontaminated according to institutional guidelines. For studies involving human samples, informed consent and institutional review board approval are required. For animal studies, institutional animal care and use committee approval is required.

Data Safety and Sharing

Proteomics data from human samples may contain identifiable information and must be handled according to applicable privacy regulations. Data sharing policies, such as the National Institutes of Health Genomic Data Sharing Policy, may apply to certain types of data. Researchers should be aware of the FAIR Guiding Principles, which emphasize that data should be Findable, Accessible, Interoperable, and Reusable. Adopting FAIR practices during data collection enables reproducible and interpretable modeling and facilitates the reuse and translational application of datasets.

Professional Escalation Criteria

Certain situations warrant escalation to a specialist or supervisor. The following criteria provide guidance.

Escalate When Sample Preparation Fails Repeatedly

If the same sample preparation failure occurs across multiple attempts despite troubleshooting, escalate to a colleague with expertise in the specific sample type or method. Persistent low peptide yield, poor digestion efficiency, or contamination may require a fundamentally different approach.

Escalate When Instrument Performance Degrades

If quality control metrics show a consistent decline in instrument performance, such as decreasing peak intensity, increasing retention time drift, or decreasing identification rates, escalate to the instrument manager or service engineer. Continuing to acquire data on a poorly performing instrument wastes time and produces unusable data.

Escalate When Data Analysis Results Are Inconsistent

If data analysis results are inconsistent across replicates or if the false discovery rate cannot be controlled at the desired threshold, escalate to a bioinformatics specialist. Inconsistent results may indicate problems with the search parameters, the protein sequence database, or the quantification method.

Escalate When Clinical or Regulatory Decisions Are Involved

If the proteomics results will be used for clinical decision-making, regulatory submissions, or other high-impact applications, escalate to a specialist with expertise in the relevant regulatory framework. Orthogonal validation by an independent method is strongly recommended, and the limitations of the bottom-up approach must be clearly communicated.

Applications Across Biological and Clinical Research

Bottom-up proteomics has been applied across a wide range of biological and clinical research areas. The following examples illustrate the diversity of applications and the specific workflow considerations for each.

Multi-Organism Systems

Studying systems containing multiple organisms requires careful attention to sample preparation and data analysis. A bottom-up proteomics workflow has been developed to study differential protein expression in fleas experimentally infected by the bacterium Bartonella henselae, the etiological agent of cat scratch disease. The workflow includes protein extraction, protein cleanup, total protein measurement, nanoLC-MS/MS data acquisition, and data analysis using Proteome Discoverer software. The protocol can be readily applied to other label-free proteomics work involving multiple proteomes from taxonomically distinct organisms. The key challenge is distinguishing peptides from each organism and correctly assigning shared peptides.

Clinical Biomarker Discovery

Serum proteomics has been used for biomarker discovery in pediatric inflammatory bowel disease. In a case-control study, serum samples from patients with inflammatory bowel disease and non-inflammatory bowel disease patients were analyzed using an aptamer-based proteomics platform measuring over 1,300 proteins. The discovery phase identified serum protein biomarkers that differentiated inflammatory bowel disease from non-inflammatory bowel disease and distinguished ulcerative colitis from Crohn's disease. Multi-protein predictors were developed and validated in independent cohorts, demonstrating the potential of proteomics for non-invasive diagnosis and subtype differentiation.

Neuroscience and Psychiatric Disorders

Proteomics of post-mortem brain tissue has identified protein changes associated with schizophrenia. In a case-control study of prefrontal cortex tissue from 96 individuals, hundreds of proteins were found to be differentially expressed between schizophrenia cases and controls. The regulated proteins included genes located in genome-wide association study loci and proteins identified by protein quantitative trait locus analysis. Gene ontology analysis revealed downregulation of mitochondrial oxidative respiration, ribosomes, and the proteasome, and upregulation of kinases and small GTPases. These findings support a role for energy deficits compromising highly ATP-dependent neuronal function.

Cellular and Molecular Biology

Bottom-up proteomics is widely used to study cellular processes such as protein secretion, extracellular matrix remodeling, and co-translational regulation. In a study of astrocytes derived from induced pluripotent stem cells from schizophrenia patients and healthy controls, mass spectrometry-based proteomics was used to profile cell lysates and secreted proteins. Compartment-specific analyses revealed that lysates were enriched for mitochondrial and nuclear pathways, whereas conditioned media were enriched for extracellular matrix and vesicle-associated proteins. Differential expression analysis revealed minimal overlap between dysregulated proteins in lysates and conditioned media, suggesting modality-specific effects of the disease-associated genetic background.

Drug Development

Automated sample preparation platforms have enabled the application of bottom-up proteomics to drug development. In a study of targeted protein degradation, an automated workflow was used to perform proteome profiling and confirm target degradation by precise protein quantification. The automated platform achieved high reproducibility and throughput, enabling the characterization of multiple compounds across multiple cell lines. This application demonstrates the importance of standardization and automation for translating proteomics into industrial and clinical settings.

Frequently Asked Questions

What is the difference between bottom-up and top-down proteomics?

Bottom-up proteomics digests proteins into peptides before mass spectrometry analysis, while top-down proteomics analyzes intact proteins. Bottom-up is the prevailing strategy because peptides are more amenable to chromatography and fragmentation, enabling high-throughput identification and quantification. Top-down provides direct information about proteoforms, including the combination of post-translational modifications on a single protein molecule, but is technically more challenging and lower throughput. Integrated approaches that combine both strategies can provide more precise proteoform characterization than either method alone.

Why is sample preparation considered the main challenge in bottom-up proteomics?

Sample preparation is laborious, prone to errors, and has historically shown low reproducibility and throughput. It is a crucial stage that affects the overall efficiency of a proteomic study. The quality of the peptides introduced into the mass spectrometer determines the quality of the data, and inconsistencies in sample preparation propagate through the entire workflow. Recent advances in automation and novel methods such as suspension trapping and bead-based digestion aim to reduce time, increase throughput, and improve reproducibility.

What is filter-aided sample preparation and when should it be used?

Filter-aided sample preparation, or FASP, is a widely used protein processing technique that uses ultrafiltration units with membranes having large molecular mass cut-offs. It allows buffer exchange, detergent removal, protein digestion, and peptide collection on a single filter device. FASP is applicable to a variety of sample types and produces high-quality peptides. It permits digestion with a variety of enzymes, allows monitoring of protein-to-peptide conversion, and enables consecutive cleavage with several proteases. It is particularly useful when detergents are required for protein solubilization but must be removed before LC-MS/MS analysis.

What is the difference between data-dependent and data-independent acquisition?

Data-dependent acquisition selects the most abundant precursor ions from each survey scan for fragmentation, which is simple and widely used but suffers from stochastic sampling of low-abundance peptides. Data-independent acquisition systematically fragments all precursor ions within defined isolation windows, providing more complete and reproducible coverage but generating complex fragment spectra that require specialized analysis software. Data-independent acquisition has emerged as a powerful technology for high-throughput, accurate, and reproducible quantitative proteomics.

How is the false discovery rate controlled in bottom-up proteomics?

The false discovery rate is controlled using a target-decoy database search strategy. The search is performed against both the target protein sequences and a set of reversed or shuffled decoy sequences. The number of decoy identifications at a given score threshold estimates the number of false target identifications, allowing the calculation of a false discovery rate. A false discovery rate of 1 percent at the peptide and protein levels is a common standard for discovery proteomics.

What are the limitations of bottom-up proteomics for studying proteoforms?

Bottom-up proteomics loses the connection between individual peptides and their parent proteoforms because proteins are digested before analysis. Peptides shared between multiple proteoforms cannot be unambiguously assigned, and the combination of post-translational modifications on a single protein molecule cannot be determined. This limitation is particularly important for studying proteoforms, where the precise combination of modifications determines function. Top-down proteomics or integrated bottom-up and top-down approaches are needed for direct proteoform characterization.

How should samples from multiple organisms be handled in bottom-up proteomics?

Samples from multiple organisms require careful attention to sample preparation and data analysis. The protein sequence database must include all expected protein sequences from all relevant taxa, and the analysis must account for the possibility of homologous peptides shared between organisms. A bottom-up proteomics workflow developed for studying fleas infected with Bartonella henselae provides step-by-step instructions for protein extraction, protein cleanup, total protein measurement, nanoLC-MS/MS data acquisition, and data analysis. The protocol can be applied to other label-free proteomics work involving multiple proteomes from taxonomically distinct organisms.

What quality control measures should be implemented in a bottom-up proteomics experiment?

Quality control measures include verifying protein quantification before digestion, monitoring digestion completeness, assessing peptide quantity and quality before LC-MS/MS, running quality control samples at regular intervals during acquisition, and monitoring instrument performance metrics over time. For quantitative experiments, equal protein loading across samples is critical. For data-independent acquisition, the quality and completeness of the spectral library directly affect the sensitivity and accuracy of the analysis.

Related Bioinformatics Guides

• Flux Balance Analysis in Metabolic Networks: Principles, Computational Advances, and Applications in Veterinary Systems Biology
• In Silico Analysis of Spike Protein Mutations and Their Impact on Host Receptor Affinity: A Computational Virology Approach
• Structural Comparison and Alignment Algorithms for Protein 3D Structures
• Normal Mode Analysis and Elastic Network Models for Protein Flexibility
• Structural and Evolutionary Analysis of Viral Entry Proteins: A Computational Approach

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.
• Bottom-Up Proteomics: Advancements in Sample Preparation.. International journal of molecular sciences, 2023.
• Comprehensive Overview of Bottom-Up Proteomics Using Mass Spectrometry.. ACS measurement science au, 2024.
• A bottom-up proteomics workflow for a system containing multiple organisms.. Rapid communications in mass spectrometry : RCM, 2023.
• Bottom-Up Proteomics Workflow for Studying Multi-organism Systems.. Methods in molecular biology (Clifton, N.J.), 2025.
• Collision energies: Optimization strategies for bottom-up proteomics.. Mass spectrometry reviews, 2023.
• Comprehensive Overview of Bottom-Up Proteomics using Mass Spectrometry.. ArXiv, 2023.
• Acquisition and Analysis of DIA-Based Proteomic Data: A Comprehensive Survey in 2023.. Molecular & cellular proteomics : MCP, 2024.
• Filter Aided Sample Preparation - A tutorial.. Analytica chimica acta, 2019.
• Enabling Next-Generation Mass Spectrometry-Based Proteomics: Standards, Proteoform Resolution, and FAIR, Reproducible, and Quantitative Analysis.. 2026.
• Co-translational profiling in the cardiac endothelium in response to LPS-induced inflammation in female mice in vivo: a proof-of-concept approach.. 2026.
• Multimodal Proteomics Reveals Dysregulated Secretion and ECM Remodelling in Schizophrenia Patient iPSC-Derived Astrocytes.. 2026.
• Serum proteomics in paediatric inflammatory bowel disease from a case-control study: biomarker discovery and ulcerative colitis-Crohn's disease differentiation.. 2026.
• Human brain prefrontal cortex proteomics identifies compromised energy metabolism and neuronal function in Schizophrenia.. 2026.
• Medicinal cannabis plant extract (NTI164) modifies epigenetic, ribosomal, and immune pathways in paediatric acute-onset neuropsychiatric syndrome.. 2026.
• A flexible end-to-end automated sample preparation workflow enables reproducible large-scale bottom-up proteomics. bioRxiv, 2025.
• Integrated top-down and bottom-up proteomics enables precise characterization of proteoforms within the protein corona. Nature Communications, 2026.

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