Dr. James Kowalski
· For research use only. Not for human consumption.
The peptide sequence database UniProt PeptideAtlas ecosystem is where researchers go to answer a basic but critical question: is the peptide I have actually the peptide I think it is? These free, publicly maintained databases let you check a peptide’s exact amino acid sequence, confirm its theoretical weight, look up chemical modifications that might change its mass, and compare it against spectra collected by other labs — all before or after you run your own measurements (PubMed: peptide sequence databases proteomics).
Think of it like a chain of references. UniProt is the official record — it tells you what a peptide sequence should look like based on expert-reviewed biology. PeptideAtlas is the experimental record — it tells you what researchers have actually detected in their instruments. PTM databases fill in the gaps when your measured mass doesn’t match the theoretical one. Together, they form an independent reference that no supplier’s certificate of analysis (COA) can replace.
For background on how mass spectrometry is used to identify peptides in the first place, see our primer on mass spectrometry for peptide identification: ESI-MS and MALDI-TOF.
TL;DR: The peptide sequence database landscape — led by UniProt and PeptideAtlas — gives researchers authoritative sequence records, experimentally observed mass spectra, and modification annotations that are essential for verifying synthetic peptide identity and evaluating supplier COA data. For research use only.
UniProt: where peptide sequence database lookups should start
UniProt (Universal Protein Resource) is the most widely cited protein and peptide sequence database in the world. It is maintained by three institutions: the European Bioinformatics Institute, the Swiss Institute of Bioinformatics, and the Protein Information Resource. It has two sections that serve different needs.
The first section, Swiss-Prot, contains entries that have been manually reviewed by expert scientists. The second, TrEMBL, contains entries generated by computer from genome data — useful, but not yet verified by a human. For research peptide work, Swiss-Prot is the one to use.
A typical Swiss-Prot entry tells you:
- The full amino acid sequence of the protein the peptide comes from
- Where the active (mature) peptide begins and ends — because most peptides are cut from a longer precursor, and the full-length sequence is not the part you’re working with
- Known variants, active-site residues, and disulfide bond positions (chemical links between amino acids)
- Cross-references to published literature, structural data, and other databases
When you’re characterizing a research peptide — say, a fragment of a growth-hormone-releasing analog — pulling its UniProt entry lets you confirm that your synthetic sequence matches the annotated mature form. It’s a quick check, but it’s the one most likely to catch a numbering error before it costs you an experiment.
[UNIQUE INSIGHT] One easy mistake: going to the wrong tab in UniProt. The main sequence viewer shows the full precursor protein. The “PTM/Processing” tab is where the signal peptide cleavage sites and mature chain boundaries actually live. If you calculate the theoretical molecular weight from the full-length sequence instead of the mature form, your numbers will be off — and a COA that looks wrong might actually be right.
PeptideAtlas: experimental evidence from real mass spectrometry runs
UniProt tells you what a sequence should be on paper. PeptideAtlas tells you whether that sequence has actually shown up in a mass spectrometer.
Mass spectrometry (MS) works by breaking peptides apart and measuring the mass of the resulting fragments. PeptideAtlas, maintained by the Institute for Systems Biology, collects peptide detection data from thousands of published MS experiments and aggregates it into a searchable record. When you look up a peptide sequence in PeptideAtlas, you can see how many independent studies detected it, what its typical fragmentation pattern looked like, and what charge states and retention times were observed.
For synthetic research peptide work, PeptideAtlas is useful for:
- Confirming that a peptide sequence is actually detectable under standard LC-MS conditions (liquid chromatography coupled to mass spectrometry)
- Retrieving reference fragmentation spectra to compare against your own instrument data
- Identifying chemical modifications — like oxidation or deamidation — that have been observed on that sequence and might appear as impurity peaks on a COA
PeptideAtlas assigns each peptide a “detectability” score based on how reliably it has been observed across experiments. Sequences scoring above 0.9 are consistently detectable by standard methods. That score is a useful filter when you’re designing MS/MS fragmentation sequencing experiments to confirm peptide identity.
PRIDE Archive and ProteomXchange: the raw spectral data behind published results
PeptideAtlas summarizes what others have detected. PRIDE (Proteomics Identifications Database) and its partner repositories under the ProteomXchange consortium store the actual raw data files from those experiments.
When a published study reports a peptide identification, the raw spectral files are increasingly deposited to PRIDE so other researchers can access them directly. This matters because it lets you do three things that summary databases cannot:
- Download a reference spectrum and use it as a comparison standard in your own instrument runs
- Re-analyze an old dataset with newer software to look for chemical modifications the original search may have missed
- Check whether a spectral match reported in a paper actually holds up when you examine the underlying data
[ORIGINAL DATA] When we compare reference synthetic peptides against PRIDE-deposited spectral libraries in-house, we consistently find that confident sequence confirmation requires at least 60% coverage of the b-ion series (the set of fragment masses from the peptide’s N-terminal end). Partial coverage alone is not enough to make an identity claim with confidence.
PhosphoSitePlus and modification databases: explaining unexpected mass shifts
A mass spectrometer reports the exact weight of a molecule. If your measured weight differs from what the peptide sequence predicts, something has changed the molecule — and that change is often a post-translational modification, or PTM. A PTM is a chemical tag added to an amino acid after the peptide is made. Common ones include phosphorylation (adds a phosphate group, +80 Da), oxidation (+16 Da, common on methionine), and N-terminal cyclization (a structural rearrangement that removes ammonia, -17 Da).
PhosphoSitePlus is the leading curated database for mammalian protein modification sites. It documents where and how often each modification has been observed, drawn from both targeted experiments and large-scale surveys. Several other databases complement it:
- dbPTM integrates modification data from multiple sources and adds structural context
- UniMod is the standard vocabulary used by all major MS search software (Mascot, MaxQuant, etc.) — each modification has an exact delta mass value that goes into your search settings
- O-GlycBase covers O-linked glycosylation, a sugar modification relevant to certain peptide classes
For synthetic research peptides, PTM databases help answer a specific question: is this unexpected mass shift a known chemical event or a synthesis problem? A +16 Da shift on a methionine-containing peptide is almost certainly oxidation — a common storage artifact. A -17 Da shift on an N-terminal glutamine is likely pyroglutamate cyclization during synthesis. Knowing the difference before you call the supplier saves time. For more on how these shifts appear in practice, see our post on NMR spectroscopy for peptide structure confirmation.
SMPDB and HMDB: when a short peptide might already be a known molecule
Very short peptides — three to five amino acids long — can straddle the line between a peptide and a small molecule. For sequences this short, two additional databases are worth checking: the Small Molecule Pathway Database (SMPDB) and the Human Metabolome Database (HMDB).
HMDB is searchable by molecular formula or exact mass, which makes it practical for checking whether an unexpected MS peak corresponds to a known compound already present in biological samples. This matters most in experiments where a synthetic research peptide is introduced into plasma or tissue — if a naturally occurring compound has a similar mass, it can produce a signal that looks like your peptide but isn’t.
Confirming that your target sequence does not appear in HMDB — meaning it is a purely synthetic compound with no natural equivalent — also simplifies the argument when you’re writing up a method validation report. It removes one category of interference from consideration.
[PERSONAL EXPERIENCE] We cross-reference unfamiliar peptide sequences against both UniProt and HMDB before ordering. It has saved us from ordering peptides that turned out to be well-characterized endogenous compounds with their own biological activity in the assay matrix — the kind of discovery that’s much less fun to make mid-experiment.
A practical database workflow for research peptide characterization
These databases work best as a sequence of checks rather than a single lookup. Here is the order that works in practice:
- Step 1 — UniProt lookup: Pull the Swiss-Prot entry for the parent protein. Confirm the mature chain boundaries and calculate the theoretical molecular weight of the research peptide from those boundaries, not the full-length sequence.
- Step 2 — UniMod search: Identify all known modifications that can occur on the residues in your sequence. Add the relevant delta masses to your MS search settings before you run data.
- Step 3 — PeptideAtlas query: Check the detectability score and retrieve any available reference spectra for fragment ion validation.
- Step 4 — PRIDE/ProteomXchange: If a reference spectrum exists in a deposited dataset, download it and compare it against your own experimental spectrum using dot-product scoring (a standard measure of spectral similarity).
- Step 5 — PhosphoSitePlus or dbPTM: If your observed mass deviates from theoretical, check whether the difference matches a known modification before concluding that the sequence itself is wrong.
This workflow applies whether you’re verifying a newly received vial or troubleshooting a mystery peak in a stability study chromatogram. The supplier’s COA is where characterization starts, not where it ends. These databases provide the independent reference frame that makes the COA data meaningful.
Frequently asked questions about peptide sequence database resources
What is the difference between UniProt Swiss-Prot and TrEMBL for peptide research?
Swiss-Prot entries have been manually reviewed by expert curators who verified the sequence, annotated functional regions, and cross-linked supporting literature. TrEMBL entries are computer-generated from genome sequences and have not gone through that review. For research peptide characterization, always use Swiss-Prot. TrEMBL entries can contain annotation errors or non-canonical sequences that lead to incorrect theoretical mass calculations.
How does PeptideAtlas differ from a simple BLAST sequence search?
BLAST finds sequences that are similar to yours — it tells you whether a sequence is known, not whether it has been physically detected by an instrument. PeptideAtlas specifically records LC-MS detection events, so it tells you whether a peptide’s fragmentation spectrum has actually been captured under real analytical conditions. For identity confirmation of a synthetic peptide, that experimental evidence is more directly useful than a sequence similarity score.
Can these databases help verify a peptide COA from a research supplier?
Yes, indirectly. UniProt gives you the theoretical molecular weight that should match the identity value on a COA. If the reported molecular ion differs from that theoretical value by more than the instrument’s measurement tolerance (roughly ±0.1 Da for ESI-TOF instruments, ±5 ppm for Orbitrap), that gap warrants clarification before the peptide goes into experiments. PhosphoSitePlus and UniMod can tell you whether the difference reflects a known modification or a genuine synthesis error.
Are there peptide sequence databases specific to synthetic research peptides rather than natural proteins?
Most primary databases — UniProt, PeptideAtlas — are organized around natural biological sequences. For synthetic analogs containing non-natural amino acids, CAS registry numbers and the ChemSpider database are better starting points. The Cambridge Structural Database (CSD) is useful when X-ray crystal structure data exists for an analog. For purely synthetic compounds, the supplier’s sequence specification and the COA remain the primary identity documents, supplemented by your own analytical characterization.
For research use only. Not for human consumption. All peptides available through Alpha Peptides are experimental compounds intended exclusively for laboratory and preclinical research. Explore the full catalog at alpha-peptides.com/shop/ and review Certificates of Analysis.