Palaeoproteomics investigates ancient protein fragments in fossils, historic artefacts and other archaeological substrates. **Finding a plausible peptide is not the same as authenticating an ancient protein or proving a species.** The gold standard teaches protein preservation, laboratory source checks, partial sequence interpretation and respectful evidence use.
Our fictional fossil tooth F01 has a small set of invented peptide matches. The real featured image is a photographed Pleistocene mammoth molar by James St. John; no peptide result is attributed to that specimen. All count-based demonstrations in this article are constructed for education.
The University of York BioArCh programme, Chemical Reviews palaeoproteomics synthesis, Nature Protocols enamel work and eduKate Gold Standard provide the professional foundations.
1. Begin with a protein claim, not an invented fossil identity
Our fictional fossil tooth F01 yields peptide-spectrum matches that may relate to enamel proteins. Some pupils immediately call it a particular mammoth species or claim to have decoded its genome. Neither follows from the initial observation. Palaeoproteomics studies preserved proteins, peptide sequences and their modifications to investigate historical and evolutionary questions. A source-appropriate conclusion depends on the tissue, preservation, peptide specificity and analytical controls, not the excitement of identifying an extinct animal. Students should first write: which protein evidence was observed, and what remains unknown about its source and age?
The University of York BioArCh centre documents applications to ancient food, material heritage, taxonomy and evolution. The photographed Pleistocene mammoth molar is a real fossil illustration, but is not F01 and has no measured proteomics data in this article. The classroom counts are entirely fictional. A strong pupil distinguishes the credited object, laboratory evidence and historical interpretation before naming a species or an ancestral relationship.
2. Distinguish amino-acid chains from ancient DNA
Proteins are amino-acid polymers expressed through cellular biology, while DNA carries nucleotide information. Certain proteins can remain informative in mineralised tissue even when recoverable ancient DNA is absent, but protein fragments represent only a limited part of an organism’s encoded sequence information. They cannot be converted directly into a whole genome or proof of language, individual cultural identity or complete biological ancestry.
The 2024 Nature Protocols review describes deep-time enamel proteins and the problem of reconstructing partial amino-acid sequences. In class, learners shade three identified peptide segments on a fictional long protein. Every blank portion remains unobserved. They should explain why a correct peptide identity can still leave most of a protein and the organism’s genetic history unknown.
3. Know which archaeological substrates preserve which proteins
Bone is rich in collagen; enamel retains a more restricted proteome with proteins such as amelogenins and enamelin. Dental calculus, historic painting binders, parchment and pottery crusts may preserve different mixtures. A collagen signal from bone answers different questions from a rare enamel sequence with variable amino-acid sites. Tissue and mineral context determine what a researcher might reasonably expect.
Provide an imaginary tooth, a bone fragment and a medieval parchment leaf. Students identify candidate protein classes and why preservation cannot be guaranteed. The Chemical Reviews palaeoproteomics synthesis covers tissues and their characteristic proteins. A strong report does not treat the phrase ‘protein recovered’ as sufficient explanation of its material origin.
4. Treat ancient peptides as fragments, not complete original proteins
Burial and time can alter ancient proteins through cleavage, oxidation, deamidation and other processes. Mass spectrometry often recovers limited peptides, not intact full-length proteins. A missing region may simply have failed to preserve or ionise; it is not proof that the organism lacked that sequence. Ancient enamel may hold some informative positions even when most sequence detail remains inaccessible.
Students compare a real published enamel-protein sequence diagram with an invented patchy one. They should identify observed and unobserved residues and avoid filling the gaps by choosing the closest reference species. This is how palaeoproteomics provides useful but incomplete biological evidence. The learning objective is a partial reconstruction whose limitations can be audited.
5. Understand mass-to-charge measurements and tandem spectra
Mass spectrometry measures ions by mass-to-charge ratio, while tandem MS can generate fragment patterns used to compare peptide sequence hypotheses against reference databases. The instrument does not directly print the ancient animal’s name. A high-scoring candidate must be evaluated using signal quality, fragment assignment and the error model. Complex backgrounds and a scarce target protein can complicate identification.
A fictional F01 report has a peak and two peptide sequences that fit part of the signal. Students ask what discriminating ions would support one. An attractive spectrum without mass calibration, reference database or uncertainty is not a taxonomic certificate. The instrument output is evidence, and the sequence interpretation is a separate analytical stage.
6. Distinguish spectrum matches from distinct peptide information
One peptide can generate repeated spectra, and several peptides can be shared among species. Forty peptide-spectrum matches do not necessarily represent forty independent taxonomic markers. Unique peptides and variable amino-acid positions are more informative for some questions than repeated observations of common collagen or enamel fragments. Researchers use quality filtering and appropriate false-discovery control.
The fictional F01 table contains forty initial matches but fourteen distinct enamel-associated peptide candidates. Students should describe why those numbers differ and why fourteen still cannot prove a unique species. Give the class a shared peptide that occurs in two hypothetical taxa; both remain plausible. The correct source record names the individual peptides and their specificity rather than simply reporting the largest possible total.
7. Account for incomplete and biased protein reference databases
A sequence-search engine can identify only sequences represented in, or inferable from, its reference collection under the chosen method. Closely related taxa share many regions, and some extinct forms have no appropriate complete reference. A database restricted to one favourite species may produce overconfident assignments. Spectral ambiguity must not be converted into certainty by naming the only available reference.
Students run a paper exercise with a peptide shared by hypothetical taxa A and B. First the reference sheet contains only A; then B is added. Did the ancient peptide change? No. The taxonomic interpretation changed when comparison data improved. This is a powerful introduction to reference bias in molecular archaeology and phylogenetic inference.
8. Treat keratin and other modern proteins as possible contamination
Modern skin, hair, laboratory materials and conservation treatments can introduce proteins into historic specimens. Keratin is a frequent laboratory background candidate. Exogenous proteins may also affect what instruments detect by competing with rare ancient peptides, even when the contaminants themselves do not appear among the final confident assignments.
A 2025–2026 Scientific Reports enamel study demonstrated this issue using controlled contamination of an ancient rhinoceros tooth. In class, T1 gives enamel-peptide evidence alongside keratin in an instrument blank. Learners should investigate controls and protein composition rather than subtract counts or publish the keratin as an ancient animal trait.
9. Use extraction, injection and independent controls correctly
Extraction blanks can reveal contamination introduced through preparation, while instrument or injection blanks address relevant analytical carryover. Separate extracts and replicated identifications may strengthen evidence, but repeated measurements of a single contaminated extract are not independent biological confirmation. A report should disclose which controls were included and what each can test.
Students evaluate two imaginary runs, one with documented extraction and injection controls and another with no control history. The guide to ancient protein studies advocates source-aware contamination and reproducibility checks. The exercise asks learners to write a statement proportionate to actual controls rather than tick an ‘authentic’ box because a device was used in a clean room.
10. Explain deamidation without inventing an amino-acid clock
Deamidation of particular amino acids and other modifications can accumulate or vary with protein preservation, chemistry and treatment. Deamidation can also occur in living organisms and under laboratory conditions; no universal fraction converts a peptide directly into an age. A modified peptide can support an ancient-origin assessment only in an appropriate wider context.
Chemical Reviews cautions that damage signatures in ancient proteins are much more context-dependent than familiar ancient DNA terminal signals. Ask pupils whether a hypothetical fragment with more deamidation must be older. It need not be. Temperature, environment, protein identity and preparation can all differ. A proper conclusion notes chemical modification as evidence with limits.
11. Assess tissue-specific compositions before selecting an exotic protein
Enamel material is expected to contain proteins compatible with enamel formation and structure under a preservation model. A reported large signal from an unrelated modern reagent deserves investigation. Yet an unexpected protein might also arise from adjacent tissue, contamination or an incomplete database. One striking assignment is not a substitute for the overall protein mixture and archaeological context.
Students compare a fictional enamel-related proteome with a sample dominated by laboratory keratin and unrelated food proteins. They propose checks using tissue anatomy, blanks, handling records and specialist validation. This teaches why the metaproteome context matters and how to avoid sensational claims of ancient diet or disease from one unexplained peak.
12. Use variable peptide sites cautiously for evolutionary inference
Amino-acid differences among related taxa can contribute to phylogenetic models, but sparse coverage and conserved regions limit resolution. A tree is a model-based hypothesis, not a direct image of historical descent. Comparing only the peptides that survived may miss distinguishing sequence positions that could change the result.
Give an invented four-taxon alignment with two observed variable sites and several unknown positions. Students identify which groupings are compatible and which remain uncertain. Nature Protocols discusses sequence reconstruction and phylogenetic validation from degraded enamel. The teacher should reward showing missing data instead of forcing one neat tree.
13. Describe deep-time discoveries without claiming they are universal
A 2025 Nature study reported enamel-protein fragments from certain mammalian fossils up to approximately eighteen million years old in the East African Rift. The work discussed sparse ancient peptides and potential evolutionary information; it did not recover entire ancient genomes or show that every fossil from that period contains similar proteins.
Students rewrite the exaggerated claim ‘DNA decoded from every eighteen-million-year-old mammal’ into an accurately sourced finding about particular enamel proteins in selected fossils. They should report which specimens, peptides and validations underlie the real article. This promotes evidence-based enthusiasm without converting an exceptional achievement into unsupported generality.
14. Connect ancient proteins to food and historic materials without conflation
Protein sequences from pottery residues, dental calculus, parchment, eggshell and other materials can support different questions about past foods, taxa and material technologies. A dietary protein’s presence does not automatically establish how often someone consumed the food, while a shared collagen fingerprint may fail to distinguish close animal species. Formation, contamination, preservation and comparison sources matter in every setting.
York BioArCh describes applications to milk consumption, historic parchment and extinct animals. Ask students which field owns each question: broad ancient protein reconstruction, rapid ZooMS taxonomic fingerprinting, or dental-calculus source analysis. The gold standard gives each technique a clear purpose and avoids turning every biomolecular find into a dramatic diet narrative.
15. Archive measurements and protect precious historic material
A useful palaeoproteomics archive preserves specimen provenance, tissue, sampling authority, processing, controls, raw spectra, peptide sequence assignments, database versions, ambiguity, modifications and downstream models. Destructive sampling of fossils and culturally significant human remains needs appropriate permissions, consultation and conservation rationale. Strong research makes reinterpretation possible without automatically demanding that rare artefacts be sampled again.
Students repair an invented report that says only ‘fossil tooth proven to be species A’. They add the spectra, observed peptides, context and alternative assignments needed for independent review. The 2018 ancient protein guidelines explicitly support full methodological reporting. A specimen may be scientifically valuable even when its species remains unresolved.
16. Prove transfer using a contested fossil peptide assignment
The final imaginary packet has a fossilised tooth, a few candidate enamelin fragments, a laboratory blank with keratin contamination and incomplete archaeological provenance. Students must separate what is measured from what is inferred, identify the best next authenticity checks and explain whether the dataset supports species identification or evolutionary placement.
A strong response may say that only limited ancient-protein candidates are supported and the fossil’s exact taxon cannot be established. It preserves the object, reports sample and laboratory uncertainties and does not invent a genetic lineage. That is the Gold Standard of Teaching Palaeoproteomics.
Worked F01 peptide audit: forty spectrum matches do not prove one species
| Invented audit field | F01 fictional result | What remains uncertain |
|---|---|---|
| Initial peptide-spectrum matches | 40 | Some refer to the same peptide |
| Distinct high-quality enamel-peptide candidates | 14 | Several may be shared by related taxa |
| Candidates replicated in independent extraction | 9 | Provenance and controls still need review |
| Keratin-like spectra in F01 | 7 | Could represent modern background |
| Keratin-like spectra in blank | 6 | Calls for contamination assessment, not direct subtraction |
| Complete full-length enamel protein sequence | None demonstrated | Missing regions cannot be invented |
This is a fictional teaching dataset, not a laboratory report about the featured mammoth. The number of peptide-spectrum matches is not the number of independent proteins, and the presence of a blank contaminant means the source of affected signals needs investigation. A tentative enamel-peptide assignment may still be informative when correctly authenticated, but its historical interpretation is constrained by sequence coverage and reference specificity.
Case A — a high peak from a common protein masquerades as decisive evidence
A fictitious report has one dominant collagen peptide and a tiny enamel peptide. The writer assigns the specimen to one extinct taxon on the basis of the dominant signal, though the collagen peptide occurs across several related animals. Ask students to identify which sequence characters would actually distinguish taxa and whether those positions were measured. A high-intensity spectral peak can provide a clear molecular observation without uniquely assigning source species.
Now add an independent enamel peptide containing a variable amino-acid position. This may improve taxonomic information, but one position may still be shared by more than one lineage or be affected by sequence ambiguity. The student should report how confidence changed and what alternative identifications remain plausible. An honest conclusion evolves with the evidence rather than using the first matching database name as a permanent label.
Case B — modern protein contamination masks older peptide information
An imagined second laboratory run on F01 has more human-keratin background and fewer confident enamel-peptide assignments. The tooth has not changed its biological identity between runs; analytical access to low-abundance signals has changed. The 2025–2026 rhinoceros enamel study documents how introduced proteins can impair recovery of ancient information even when the introduced species is not directly identified.
Ask which controls and data would help distinguish genuine loss of ancient protein from instrumental suppression or inconsistent sample preparation. A valid response keeps the archaeological specimen, experimental preparation and spectrum-identification stages separate. It does not claim that modern proteins prove the fossil was handled by a specific living person.
Case C — deamidation is not a stopwatch
Two fictional protein preparations contain different proportions of deamidated peptide candidates. A pupil interprets the sample with more modification as necessarily older. But chemical conditions, protein identity, preservation and preparation may differ. Ancient protein modification is useful evidence only as part of a wider source-authentication assessment.
Students should propose contextual checks: source tissue, burial chemistry, laboratory processing, negative controls and independent archaeological dates. The appropriate publication may describe a pattern of chemical degradation consistent with ancient material while declining an exact numeric age. This distinction prevents confusion with radiocarbon or other dedicated dating methods.
Case D — sparse alignment cannot yield a unique evolutionary tree
An invented four-taxon sequence alignment has only two securely observed variable positions from F01, with several positions unknown. Two different evolutionary arrangements are compatible with the limited evidence. Ask whether a diagram showing one branch arrangement proves the alternative impossible. It does not.
Students should compare the observed characters, the taxa represented in the reference and the missing positions. They can report a range of compatible interpretations and ask for more peptide evidence or suitable reference sequences. The exercise teaches why a palaeoproteomic tree must carry a model and uncertainty, not only coloured branches.
Case E — fossil preservation differs across tissues and sites
One fossilised mammal tooth preserves enamel peptides while another similar-looking tooth from a different burial environment yields no informative protein sequence. The appearance of a fossil alone does not explain the result. Temperature, mineral structure, water and chemical history can affect preservation, and the absence of signal is not proof the animal lacked those proteins during life.
Ask the class what source metadata and analytical controls could distinguish true non-preservation from laboratory detection failure. A negative result can be valuable if its method is documented. The teacher should discourage selecting only successful samples for publication, because survival bias can distort impressions about species and time periods.
Case F — historical material and culinary evidence belong to different questions
A parchment sample yields a collagen fingerprint and a pottery crust yields food-related peptide candidates. The first may support a source-animal taxon under suitable ZooMS reference data; the second may help investigate ingredients or processing, with contamination and context constraints. Neither alone reveals the entire manuscript’s production history or one person’s usual diet.
Students make a source card for each, explaining the substrate, expected protein, method and historical question. They then identify a suitable complementary evidence source: manuscript dating and documentary provenance for parchment, or pottery context and organic residues for food. This prevents ancient protein analysis from turning every biomolecular signal into a complete narrative of human practice.
A public museum caption should not pretend the genome was decoded
A research team identifies several authentic enamel peptides in an old fossil and publishes an age supported by independent geology. The museum caption says ‘scientists read the entire genome from this fossil’. That is wrong. Students should revise it to say that a limited set of ancient protein fragments was recovered and used for cautiously stated comparisons under the original study’s methods.
The exercise tests scientific communication: can the pupil retain what makes the finding important without inventing completeness? They should include the source collection, evidence type and key uncertainty. The result is more informative to the public than a dramatic headline that cannot be supported by the published data.
Four-lesson learning sequence
- Lesson 1: distinguish proteins from DNA and identify likely sources in bone, enamel and archaeological materials.
- Lesson 2: read peptide mass spectra conceptually and separate repeated spectrum matches from distinct protein sequence evidence.
- Lesson 3: evaluate contamination, chemical alteration, reference bias and limited phylogenetic inference.
- Lesson 4: analyse a new fictional fossil peptide packet, retaining every uncertainty and a responsible ethical sampling rationale.
An advanced evidence card: count distinct sequences before drawing a tree
Students sometimes encounter a molecular dataset with many confident spectra and assume that the organism’s evolutionary position is equally confident. Give the class an invented alignment of five taxa in which a peptide of twenty amino acids is identical in every taxon, while a second peptide has one distinguishing residue shared by only two taxa. The first peptide may provide useful confidence that a particular protein family is present, but it cannot distinguish the five candidate species. The second provides some comparative information, but if that residue is susceptible to chemical alteration or lies near an uncertain identification, the taxonomic conclusion is weaker than a clean table might suggest. Ask students which claims remain stable when the uncertain peptide is excluded.
Next replace the fossil’s missing residues with the letters from a favoured reference species and draw a phylogenetic tree. The resulting branches may look more resolved, but the added information came from the reference, not the fossil. Pupils should identify that mistake explicitly: missing observations cannot be transformed into measured evidence by reference filling without a justified modelling step and transparent uncertainty. The best published report retains the incomplete alignment, specifies which variant positions were directly detected, and presents alternative relationships where the data cannot discriminate them.
Contamination triage: no single negative control answers every question
An instrument blank and an extraction blank test different parts of the workflow. A specimen could show no keratin in the injection blank but still carry exogenous material introduced during excavation decades earlier. Conversely, a contaminant appearing in a blank can undermine confidence in a peptide assignment even when the fossil’s archaeological provenance is excellent. Ask pupils to trace a hypothetical peptide through museum storage, sampling, extraction, instrument acquisition and database search. For each stage they identify what a credible control can reveal and what it cannot. The method discourages false binary statements such as ‘one blank was clear, therefore all ancient proteins are authentic’.
The teacher should also discuss conservation: obtaining additional samples is not always the right answer to weak original documentation. If a rare human or extinct-animal fossil has already been sampled, a stronger first step may be re-examination of archived spectra, instrument logs, existing blanks and the proteomic search settings. A good research culture protects irreplaceable objects and makes older data reusable. The experiment is successful when a student can explain a defensible verification pathway without promising that every ambiguity will be resolved by more destructive sampling.
Frequently asked questions
Can proteins survive after recoverable ancient DNA disappears?
Sometimes. Survival depends on tissue and environment, and the protein data do not equal a complete genome.
Does a peptide-spectrum match identify the fossil species?
Not necessarily. Shared peptide sequences and reference limitations may prevent a unique taxonomic assignment.
Can deamidation provide the fossil’s age?
Not as a universal reliable clock. It depends on source, environment and laboratory processing.
What distinguishes palaeoproteomics from ZooMS?
Palaeoproteomics broadly studies ancient protein preservation and sequences; ZooMS focuses on collagen-peptide mass-fingerprint taxonomic identification.
Palaeoproteomics sources and related eduKate research
- University of York BioArCh: Palaeoproteomics and ZooMS
- Chemical Reviews: Palaeoproteomics
- Nature Protocols: Enamel Protein Sequence Reconstruction
- Nature: 18-Million-Year-Old Enamel Proteomes
- Scientific Reports: Contamination and Enamel Protein Detection
- Nature Ecology & Evolution: Guide to Ancient Protein Studies
- Mammoth fossil molar reference image
- Teaching ZooMS
- Teaching Ancient DNA Analysis
- Teaching Dental Calculus Analysis
- Gold Standard master hub
The gold standard of teaching Palaeoproteomics is to read ancient peptide evidence accurately, preserve missing information and contamination checks, and make biological or historical claims no stronger than the actual protein sequence and archaeological context permit.