- Peptide-spectrum matching (PSM) is a fundamental computational technique in proteomics and bioinformatics used to identify peptide sequences from tandem mass spectrometry (MS/MS) data. It involves comparing experimentally measured fragmentation spectra with theoretical spectra predicted from candidate peptide sequences. The objective is to determine which peptide sequence most closely explains the observed mass-to-charge ratios and intensities of fragment ions. PSM is a central step in protein identification, quantitative proteomics, post-translational modification analysis, and large-scale characterization of biological samples.
- Peptide-spectrum matching is commonly performed using liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS). In a typical workflow, proteins are extracted from cells, tissues, microorganisms, or biological fluids and digested into peptides using enzymes such as trypsin. The peptide mixture is separated by liquid chromatography and introduced into the mass spectrometer through an ionization method such as electrospray ionization (ESI). During the first stage of mass spectrometry (MS1), the instrument measures the mass-to-charge ratios (m/z) of peptide precursor ions. Selected precursor ions are isolated and fragmented through collision-induced dissociation (CID), higher-energy collisional dissociation (HCD), electron-transfer dissociation (ETD), or other fragmentation methods. The resulting MS/MS spectra contain fragment-ion information that can be used to determine peptide sequences.
- The theoretical basis of PSM depends on peptide fragmentation. When peptide bonds break during tandem mass spectrometry, characteristic fragment ions are produced. In collision-based fragmentation, b-ions generally retain the N-terminal portion of the peptide, whereas y-ions retain the C-terminal portion. The mass differences between consecutive fragment ions can reveal the amino acid residues in the peptide sequence. For example, the difference between two adjacent b-ions or y-ions may correspond to the mass of a particular amino acid residue. A series of matching fragment ions therefore provides evidence that a candidate peptide sequence is consistent with the experimental spectrum.
- The first major step in peptide-spectrum matching is candidate peptide generation. Search software uses a protein sequence database, such as UniProt, RefSeq, or Ensembl, to generate theoretical peptide sequences through in silico digestion. The digestion rules are selected according to the experimental enzyme, commonly trypsin. The software calculates the theoretical molecular masses of the resulting peptides and predicts their possible fragment ions. Candidate peptides are then filtered according to precursor mass, charge state, peptide length, enzyme specificity, and other search parameters. This reduces the number of possible sequences that must be compared with each experimental spectrum.
- Precursor mass matching is an important filtering step. The measured precursor mass provides an estimate of the molecular mass of the peptide, while its mass-to-charge ratio depends on the number of charges carried by the ion. The software calculates the theoretical mass of each candidate peptide and compares it with the experimental precursor mass. The difference between observed and theoretical mass is called mass error and is commonly expressed in parts per million (ppm) or daltons (Da). High-resolution instruments, such as Orbitrap and Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometers, provide accurate mass measurements that help reduce the number of candidate peptides.
- After candidate generation, theoretical fragment-ion masses are compared with the experimentally observed MS/MS peaks. The algorithm determines whether predicted b-ions, y-ions, and other relevant fragment ions are present in the spectrum within an acceptable mass tolerance. It may also evaluate fragment-ion intensities, the number of matched peaks, sequence coverage, precursor mass accuracy, and the distribution of matched ions along the peptide sequence. A good match generally contains multiple sequence-informative fragment ions that support the candidate peptide rather than relying on a few isolated peaks.
- Different search engines use different algorithms and scoring systems to evaluate peptide-spectrum matches. Mascot uses probability-based scoring to estimate the significance of observed matches. SEQUEST uses cross-correlation between experimental and theoretical spectra, while Comet provides an efficient search implementation based on a similar approach. X!Tandem uses probabilistic methods to evaluate peptide assignments. MS-GF+ applies statistical modeling to calculate peptide-spectrum match scores and associated significance estimates. Andromeda, integrated into MaxQuant, supports large-scale peptide identification, while MSFragger uses a fast fragment-ion indexing approach that supports conventional and open modification searches. Proteome Discoverer provides an integrated platform for processing mass spectrometry data using different search engines and analysis workflows.
- PSM scoring is a central component of peptide identification. A scoring algorithm assigns a numerical value to each candidate match based on how well its predicted spectrum agrees with the experimental spectrum. The score may incorporate matched fragment-ion peaks, mass accuracy, ion intensity, sequence-specific evidence, and other characteristics. Higher scores generally indicate stronger agreement, but scores from different search engines are not necessarily directly comparable because they use different mathematical models. Statistical significance, estimated error rates, and additional validation are needed before accepting a peptide identification.
- Search parameters strongly influence PSM results. Important parameters include precursor mass tolerance, fragment mass tolerance, enzyme specificity, missed cleavages, minimum and maximum peptide lengths, charge states, and chemical modifications. Fixed modifications are assumed to occur consistently at specified amino acid residues, whereas variable modifications are allowed to occur at selected positions. Carbamidomethylation of cysteine is commonly treated as a fixed modification when cysteine alkylation is performed during sample preparation. Oxidation of methionine and acetylation of protein N-termini are common variable modifications. Phosphorylation, methylation, glycosylation, ubiquitination, and other post-translational modifications may also be included when relevant. Excessive variable modifications increase the number of candidate peptides and computational complexity, so search settings should reflect the actual experiment.
- Peptide-spectrum matching can be performed using conventional database searching, open searching, spectral library searching, and predicted-spectrum approaches. Conventional database searching compares experimental spectra with theoretical spectra generated from known protein sequences and a defined set of modifications. Open searching allows a broader range of mass shifts and can help identify unexpected modifications or chemical changes. Spectral library searching compares experimental spectra with previously observed reference spectra. Predicted-spectrum approaches use computational models to generate expected fragmentation patterns for candidate peptides, allowing identification even when an experimental spectral library is unavailable.
- False discovery rate (FDR) estimation is essential because large-scale PSM analyses inevitably generate incorrect matches. Target-decoy searching is widely used to estimate the proportion of false identifications. In this approach, experimental spectra are searched against genuine target protein sequences and artificial decoy sequences. Matches to decoy sequences provide an estimate of random or incorrect identifications among the accepted target matches. FDR is commonly controlled at the peptide-spectrum match, peptide, and protein levels, often using a threshold of 1%. The exact threshold and reporting strategy should be appropriate for the experimental design and intended application.
- A commonly used expression for false discovery rate is:
- [
\mathrm{FDR}=\frac{\text{Estimated number of false identifications}}{\text{Total number of accepted identifications}}\times 100
] - Plain-text formula – easy to copy:
- FDR (%) = (Estimated number of false identifications / Total number of accepted identifications) × 100
- Here, estimated false identifications represent the number of incorrect matches inferred through a statistical procedure, such as target-decoy searching. Total accepted identifications refers to the number of peptide-spectrum matches, peptides, or proteins retained after applying the selected confidence threshold. The calculation must be interpreted according to the particular target-decoy method and FDR estimation procedure used by the software.
- Peptide-spectrum matching is also important for identifying post-translationally modified peptides. Chemical modifications alter peptide masses and may influence fragmentation patterns. Search engines can account for these changes by including specified modifications in their theoretical peptide models. For example, phosphorylation adds a phosphate group to serine, threonine, or tyrosine residues, while acetylation and methylation produce other characteristic mass shifts. Modified peptide identification may require enrichment methods such as immobilized metal affinity chromatography (IMAC) or titanium dioxide (TiO₂) enrichment for phosphopeptides. Correct localization of a modification requires fragment-ion evidence that distinguishes between possible modified residues.
- Several experimental factors influence PSM quality. Incomplete fragmentation, low peptide abundance, co-isolation of multiple precursor ions, ion suppression, chemical noise, and overlapping isotope distributions can produce complex or incomplete spectra. The fragmentation method and collision energy influence which ions are generated and their relative intensities. Instrument resolution, mass accuracy, isolation width, and scan speed also affect the quality of spectral information. Sample preparation, digestion efficiency, chromatographic separation, and contamination can further influence the reliability of peptide-spectrum matching.
- PSM is performed differently in data-dependent acquisition (DDA) and data-independent acquisition (DIA). In DDA, selected precursor ions are fragmented individually, producing MS/MS spectra that are generally easier to interpret and match with theoretical spectra. However, precursor selection may be stochastic, resulting in inconsistent sampling of low-abundance peptides. In DIA, groups of precursor ions are fragmented within predefined mass-to-charge windows. The resulting spectra contain fragment ions from multiple peptides, making interpretation more complex. Specialized computational algorithms separate and analyze these overlapping signals using spectral libraries or predicted fragmentation patterns. DIA software includes DIA-NN, Spectronaut, EncyclopeDIA, Skyline, and OpenSWATH.
- Peptide-spectrum matching results contribute directly to protein inference. After peptides have been identified, their sequences are mapped to protein databases to determine which proteins are supported by the identified peptides. Unique peptides provide evidence for specific proteins, while shared peptides may correspond to multiple homologous proteins or isoforms. Protein inference algorithms combine peptide-level evidence to assign protein groups and estimate identification confidence. Reliable PSM is therefore essential because incorrect peptide assignments can lead to incorrect protein identifications and misleading biological conclusions.
- PSM also provides the foundation for many quantitative proteomics workflows. Identified peptide signals can be quantified through MS1 peak intensities, fragment-ion intensities, spectral counts, or reporter ions from isobaric labels such as tandem mass tags (TMT) and iTRAQ. Label-free quantification, stable isotope labelling by amino acids in cell culture (SILAC), and targeted methods such as selected reaction monitoring (SRM), multiple reaction monitoring (MRM), and parallel reaction monitoring (PRM) all depend on accurate peptide identification. In targeted proteomics, predefined peptide sequences and their characteristic fragment ions are monitored to provide sensitive and reproducible measurements.
- Peptide-spectrum matching has applications in biomedical research, pharmaceutical development, microbiology, agriculture, and animal breeding. In medicine, it supports the identification of disease-associated proteins, biomarkers, and modified peptides. In pharmaceutical research, it helps investigate drug targets, molecular responses, and potential off-target effects. In microbiology, it contributes to microbial protein characterization, pathogen identification, and studies of antimicrobial resistance. In animal science, PSM supports the investigation of proteins associated with growth, feed efficiency, milk composition, meat quality, egg production, fertility, immune responses, and environmental adaptation. It is also useful for comparing proteomes across breeds, species, tissues, and developmental stages.
- Despite its importance, peptide-spectrum matching has several limitations. Similar peptide sequences may produce comparable fragmentation patterns, while incomplete or noisy spectra may not contain enough information for confident identification. Incomplete reference databases can prevent the identification of genuine peptides, particularly in non-model organisms and genetically diverse populations. Unexpected modifications and sequence variants may also lead to incorrect or missed matches. Differences between search engines, scoring methods, and parameter settings can produce different results from the same experimental data. Appropriate database selection, FDR control, quality assessment, biological replication, and reproducible computational workflows are therefore essential.
- Overall, peptide-spectrum matching connects experimental tandem mass spectrometry with computational peptide sequence analysis. By comparing observed fragmentation spectra with theoretical or reference spectra, PSM algorithms identify peptide sequences and provide the evidence required for protein identification, post-translational modification analysis, and quantitative proteomics. It is consequently one of the most important analytical steps in modern mass spectrometry-based proteomics and bioinformatics.