Matthew Collins
LightNews LightNews
banner
matthewcollins.bsky.social
Matthew Collins
@matthewcollins.bsky.social
4.5K followers 5K following 540 posts
Archaeology, University of Cambridge | The Globe, University of Copenhagen
Ancient Proteins | Medieval Manuscripts | Proteomics and AI | 🇺🇦
Posts Replies Media Videos
matthewcollins.bsky.social
Thanks to @isbarchaeology.bsky.social and @isba11.bsky.social for a great conference.

I mentioned the legacy of Henk Kars of VU who established ISBA

Geoarchaeology closed. Here is the link to the petition to save Earth Sciences @vuamsterdam.bsky.social

www.change.org/p/red-aardwe...
More than 500 international scholars attended ISBA 11. Here are many of them looking up at the camera in a large group shot
September 1, 2025 at 4:05 PM
matthewcollins.bsky.social
🧵 Thread: 1️⃣/ Caroline Tokarski delivering a masterclass on how #massspec revolutionises our understanding of cultural materials! Her slide perfectly encapsulates the multifaceted insights we can extract - from artist intention to societal context. #teammasspec #heritage
Caroline in a patterned dress presents to an audience in a tiered lecture theatre with curved wooden seating. The projection screen shows a colorful diagram with molecular structures and text about "MATERIALS (ID, Modification)" including bullet points about understanding artist intention, artwork changes, treatment effects, and societal context. Attendees sit on wooden benches taking notes, with laptops open throughout the amphitheatre.
June 20, 2025 at 10:42 AM
matthewcollins.bsky.social
Alba Alvarez Martin from @rijksmuseum.bsky.social
presenting "Making Mass Spectrometry Accessible in Cultural Institutions".

The historical irony isn't lost as Rembrandt's anatomy lesson is juxtaposed with #teammasspec in an amphitheatre designed for anatomical dissection.
A packed amphitheatre lecture hall with curved wooden tiered seating filled with attendees listening to a presentation. The projection screen displays a split image comparing modern mass spectrometry equipment on the left with Rembrandt's famous 1632 painting "The Anatomy Lesson of Dr. Nicolaes Tulp" on the right, branded with "RIJKS MUSEUM". The historic venue features elegant wooden architecture with multiple levels of curved benches rising steeply upward.
June 20, 2025 at 10:33 AM
matthewcollins.bsky.social
Alternative perspectives at the "Advanced Mass Spectrometry in Cultural Heritage" symposium in Bordeaux!

www.u-bordeaux.fr/evenements/s...
The 3rd of these essential gathering brings together experts from across the globe to explore cutting-edge applications in heritage science. #teammasspec
Bird's eye view of a circular lecture theatre with curved wooden tiered seating arranged in concentric arcs. At the bottom, a projection screen displays "Advanced mass spectrometry applied to cultural heritage Symposium" with University of Bordeaux branding. Two people stand near the presentation area at the front, with wooden chairs and desks arranged in the curved amphitheatre design typical of academic venues. A woman with glasses and light blue striped shirt sits at a curved wooden desk in a tiered lecture theatre. The amphitheatre features multiple levels of curved wooden benches rising upward, with another person visible in the distant upper seating. The architectural design shows elegant curved lines and wood paneling throughout the academic venue.
June 20, 2025 at 10:10 AM
matthewcollins.bsky.social
Caroline Tokarski kicks off the 2025 Bordeaux Proteomics Summer School: Advanced Mass Spectrometry Applied ro Cultural Heritage.

So happy to be here and really looking forward to this!
A lecture theatre with a female professor explaining the outline of the course.
June 16, 2025 at 7:24 AM
matthewcollins.bsky.social
Håkonshallen seems to be an appropriate venue to kick off Codicum 😀 our @erc.europa.eu SYNERGY grant

www.codicum.eu
May 14, 2025 at 3:56 AM
matthewcollins.bsky.social
Professor Anna Källén, Umeå University, speaking around her new book The Trouble with Ancient DNA: Telling Stories of the Ancient Past with Genomic Science (The University of Chicago Press, 2025)
@ucph-soa.bsky.social
Tobias Richter introduces Anna Källén Anna Källén introduces her talk.
May 13, 2025 at 3:14 PM
matthewcollins.bsky.social
3️⃣ Method involves minimally destructive sampling - a huge advantage over traditional approaches. The technique extracted and identified sex-linked peptides from third molars with surprising consistency. #palaeoproteomics #archaeologicalscience

A 'why did I not think of that moment?' 🤔
Orthogonal partial least squares discriminant analysis scores plot demonstrating separation between male and female dental enamel samples.
May 6, 2025 at 3:12 AM
matthewcollins.bsky.social
🧵 Thread: 1️⃣ Fascinating paper by Stewart et al. shows robust method for sex determination using peptides in tooth enamel.

Great potential for #archaeology with fragmentary remains where DNA is degraded. #palaeoproteomics #ZooMS #teammasspec
figshare.com/s/a861c0d74d...
Diagram illustrating sex determination using amelogenin proteins in tooth enamel. The top shows X and Y chromosomally encoded protein isoforms (AMELX in blue, AMELY in red) with their amino acid sequences highlighted. The middle section displays amelogenin's structural progression from protein to nanospheres to matrix to enamel rods, alongside images of developing tooth structure with labeled components (ameloblasts, enamel, dentin, odontoblasts, pulp). The bottom shows Raman spectroscopy setup with IR laser and detector analyzing a tooth, demonstrating how incident light produces both Rayleigh scattering and Raman scattering from molecular vibrations, resulting in sex-specific spectra shown in a graph (blue and red lines) with male and female symbols. Biological sex determination is essential for analyzing human skeletal remains in archaeology, anthropology, and forensic science. This study investigates whether Raman spectroscopy of intact human dental enamel can be used as a non-destructive method for sex estimation. Orthogonal partial least squares discriminant analysis (OPLS-DA) and logistic regression identified sex-specific spectral characteristics in 88 human teeth from 47 modern individuals (26 females, 21 males). The OPLS-DA model showed excellent performance, with R2Y(cum) = 0.943 and Q2Y(cum) = 0.895. Raman shift wavenumbers at 373, 1182, and 1600 cm−1 were identified as the most reliable discriminators and included in a final logistic regression model. This model achieved an area under the receiver operating characteristic curve (ROC-AUC) of 0.98, a sensitivity of 0.87, and a specificity of 0.94. Our results indicate that Raman spectroscopy can effectively differentiate male and female human dental enamel based on subtle differences in molecular composition—possibly linked to residual differences between AMELX and AMELY-derived proteins.
May 6, 2025 at 3:12 AM
matthewcollins.bsky.social
The extended deadline for abstract submission for the next #ICAZ working group meeting on #Archaeozoology, Genetics, Proteomics and Morphometrics (AGPM) is Friday, 16th of May.

You have just over two weeks!

sites.google.com/palaeome.org...

sites.google.com/palaeome.org...
April 30, 2025 at 6:02 PM
matthewcollins.bsky.social
For UK-based writers, statements from

@societyofauthors.bsky.social
societyofauthors.org/2025/03/21/t...

@alcs.co.uk
www.alcs.co.uk/news/alcs-co...

on illegal (in UK) use by #Meta of the #LibGen dataset

Search LibGen from the @theatlantic.com article here

www.theatlantic.com/technology/a...
Screen shot of image of a seach from the LibGen library showing an example of work by Mike Buckley and Matthew Collins 2011 Antique Collagen survival and its use for species identification in Holocene-lower Pleistocene bone fragments from British archaeological and paleontological sites https://www.pagepress.org/journals/antiqua/article/view/antiqua.2011.e1 https://doi.org/10.4081/antiqua.2011.e1 Proteins have long been known to persist in Quaternary bone fossils and are often targeted as a source of carbon used in radiocarbon dating and stable isotope analyses for determining provenance and obtaining dietary information. We have previously reported a technique using the dominant structural protein collagen (type I) as a source of genetic information for species identification in modern and relatively young (Holocene) archaeological samples. We report a systematic investigation of amino acid composition and collagen peptide mass fingerprints (PMF), for a range of samples dating back approximately 1.5 million years. Extrapolation from high temperature experimental decomposition rates predict that at a constant 10°C (the approximate mean annual air temperature in Britain today) it will take between 0.2 and 0.7 Ma for levels of collagen to fall to 1% of their original concentration in an optimal burial environment. Even when the glacial intervals of the British Quaternary are factored into the temperature calculations, the more conservative of these two estimates extends the range for collagen sequencing to the Lower Pleistocene as confirmed by the presence of collagen peptides in bones from the Weybourne Crag (~1.5 Ma). Collagen fingerprinting can extend the range of identifiable taxa present at sites with large assemblages of fragmentary bone material such as that encountered at the ~900 Ka site at Happisburgh (Norfolk, UK) recently identified as showing signs of the earliest humans in Britain.
April 18, 2025 at 6:23 AM
matthewcollins.bsky.social
New review paper by @mirandaevans.bsky.social
explores how #palaeoproteomicis is revolutionizing our understanding of how ancient people cooked 🏺

Future directions include developing minimally invasive sampling methods, better authentication techniques, and understanding preservation.
A typical shotgun proteomic pipeline including protein extraction, analysis by liquid chromatography tandem mass spectrometry, and data analysis
April 9, 2025 at 6:27 PM
matthewcollins.bsky.social
4/🧵 These books offer a stunning window into medieval globalization. Bookbinding materials weren't just local - they came from thousands of miles away, through complex trade routes that linked Arctic hunters to mainland European monasteries.
Figure 7. Map showing the medieval trading routes (from Nicholas D, Pounds NJG. 1977 An economic history of medieval Europe. Rev. Belg. Philol. Hist. 55, 963–964 and Bautier R. 1952 Les principales étapes du développement des foires de Champagne. Comptes Rendus de Séances Academiques Inscriptions BellesLettres 96, 314–326)
April 9, 2025 at 6:04 AM
matthewcollins.bsky.social
3/🧵 Why did inland French monks use SEAL skins? The colour?

Cistercians valued white clothing and objects to distinguish themselves from other sects.

Dyed cloth was the epitome of luxury in the early Christian world. Seal pup skins would've appeared 'white'— perfect for their aesthetic.
A Young Camaldolese or Cistercian Monk 1506 - 1508. Public domain photograph of 16th-century portrait painting,
April 9, 2025 at 6:04 AM
matthewcollins.bsky.social
2/🧵 The big revelation? DNA analysis traced these skins to harbor seals from Scandinavia and Scotland, harp seals from Iceland/Greenland, and even a bearded seal (a strictly Arctic species)! This shows Cistercian monks were connected to extensive Norse trade networks.
a phylogenetic tree of 54 harp seal mitogenomes, including the ancient sample EL58 and 53 modern individuals from geographically distinct populations across the North Atlantic and Arctic. Coloured circles at terminal nodes represent geographic origins, such as the Southern Gulf, Newfoundland and Labrador (including ice front and replicate), Greenland Sea, and White Sea. The EL58 sample is positioned within clade A, specifically nested in subclade A2, indicating a close relationship with modern seals from the Southern Gulf and Greenland Sea. Bootstrap values at nodes indicate the robustness of inferred relationships. This figure is adapted from Carr et al. (2015), who applied quantitative phylogenomic analysis to investigate the spatial structure of harp seal populations using complete mitochondrial genomes. Carr SM, Duggan AT, Stenson GB, Marshall HD. 2015 Quantitative phylogenomics of within-species mitogenome variation: Monte Carlo andnon-parametric analysis of phylogeographic structure among discrete transatlantic breeding areas of harp seals (Pagophilus groenlandicus).PLoS One (ed. A Janke), 10, e0134207. (doi:10.1371/journal.pone.0134207) The likely geographical origin of harbour seal skins covering the bindings (EL52, EL53, EL54 and EL55; marked in yellow) was inferred by placing into a median-joining haplotype network with a large sample of mtDNA haplotypes from contemporary North Atlantic harbour seals. Black dots denote missing haplotypes and marks on branches are mutations.
April 9, 2025 at 6:04 AM
matthewcollins.bsky.social
🧵 Is this a Dire wolf... 🤔?

No...

Our @erc.europa.eu Beasts to Craft team led by Élodie Lévêque uncovered the true identity of mysterious hairy covers on #medieval #Cistercian #manuscripts - they're ...
👇
Image fo a seal - by someone that has not seen one - from Valenciennes, ms 0320, Thomas de Cantimpré, Albert le Grand. Liber de natura rerum, vers 1280, 230 × 156 mm. Folio 116, Image : IRHT https://patrimoine-numerique.ville-valenciennes.fr/ark:/29755/B_596066101_ms_0320 (http://initiale.irht.cnrs.fr/codex/7876). The original colour of the skin (before it degraded) - but what is it? (image Elodie Leveque) Mock-up of the original book binding by Elodie Leveque. Note that the fur is too dark in this mockup (image Elodie Leveque). In 2016, the team in 2016 first confronted the collection at Troyes - but what are the bindings made of? Image of five people, one sitting, discussing a large bound medieval text held by supports in a modern library.
April 9, 2025 at 6:04 AM
matthewcollins.bsky.social
2/4 About CODICUM: We're an @erc.europa.eu funded project studying medieval manuscripts (1000-1500) using scientific analysis and digital techniques to reconstruct lost books and map knowledge networks.

sites.google.com/palaeome.org...
An image of the four members of the Codicum team examining a manuscript fragment. Three are sitting and one is standing. In the background are books from the library.
April 6, 2025 at 8:38 PM
matthewcollins.bsky.social
Your comment

'MIAPE for proteomics (www.nature.com/articles/nbt...) ...published in 2007 and needs updating' hits the nail on the head.

Re-analysis of PRIDE highlights insufficient metadata on preparation steps that impact observed degradation patterns.

ALT text summarises your LinkedIn post
Major Problems Identified Insufficient Methods Reporting Missing details on data processing, normalization, outlier handling, and reference databases. Poor Data Repository Practices Unannotated or incorrectly uploaded data (raw vs. processed). Lack of accompanying metadata/annotation files. Lack of Code/Software Transparency Custom scripts not shared (e.g., on GitHub). Software versions and parameters not specified. Root Causes (From Comments) Trainee-Driven Workflows: PhD students/postdocs often lack training in reproducible practices. Data Loss: Labs frequently lose raw/processed data due to poor archiving. PI Knowledge Gaps: PIs without computational backgrounds may not enforce standards. Publish-or-Perish Culture: Speed prioritized over rigor. Proposed Solutions Short-Term Actions Institutional Training: Graduate courses on data management, bioinformatics SIGs. Community Resources: YouTube tutorials, Coursera courses, or journal-hosted webinars. Standardized Reporting: Updated guidelines (e.g., MIAPE for proteomics https://www.nature.com/articles/nbt1329 ) and cross-omics consensus frameworks. Long-Term Shifts Mandate Code/Data Sharing: Journals and funders could enforce stricter deposition policies. Adopt Best Practices: Version control (Git), containerization (Docker), and detailed READMEs. Standardized annotation and backup protocols.
April 6, 2025 at 8:49 AM
matthewcollins.bsky.social
doi.org/10.1016/j.fo...

This study aims to evaluate the different 'cooking' methods on the quality of Basa (Pangasius) fish 🐟 bladder collagen - some of which exhibit more disruption of fibrils.

It could be relevant to #palaeoproteomics 🤔

Seriously @elsevierconnect.bsky.social #openscience?
Data availability The data that has been used is confidential.
March 23, 2025 at 6:11 AM
matthewcollins.bsky.social
4️⃣ As someone who has watched our field evolve, I'm particularly interested in seeing new methodological advances and cross-disciplinary applications. The Copenhagen venue will provide excellent facilities for these discussions.

sites.google.com/palaeome.org...
Image inside the Humanities building My favourite bike sheds at the campus.
March 17, 2025 at 11:21 PM
matthewcollins.bsky.social
2️⃣ Opening keynotes by Jean-Denis Vigne (CNRS/MNHN Paris) and Prof. Eline Lorenzen (Globe Institute, Copenhagen), with closing keynote from Dr. Mary Prendergast (Rice University). An excellent lineup of speakers! #zooarchaeology #ancientDNA
Jean-Denis Vigne (CNRS/MNHN Paris) Prof. Eline Lorenzen (Globe Institute, Copenhagen) Dr. Mary Prendergast (Rice University)
March 17, 2025 at 11:21 PM
matthewcollins.bsky.social
There's a new doctor in the house!

Impressive defence by Dr @yc386.bsky.social of her work advancing computational proteomics for #palaeoproteomics.

The team had ideas on who should aid her to explore the dark proteome!
Laboratory staff celebrating with a LEGO Darth Vader set. One person in blue holds up the box, while three colleagues (a woman in green top, a man in brown jacket with glasses, and a man in grey) stand nearby with shopping bags including a yellow LEGO bag. The scene appears to be in a lab or office space with white walls and coat hooks visible in the background.
March 16, 2025 at 9:15 AM
matthewcollins.bsky.social
5/9 Most impressive is the tool's ability to identify taxonomic mixtures in samples like coprolites, contaminated specimens, and to suggest taxonomies not represented in sequence databases (including extinct taxa). #archaeology
Overview of the ClassiCOL pipeline. (A) Schematic of the ClassiCOL algorithm workflow starting from a search engine output file. The numbers represented in the table are randomly chosen to highlight the possible difference between two taxonomic branches. (B). Schematic overview of peptide reuse from independent mutations after the speciation event, which are important for classification over different taxonomic levels. The peptides represented in the scheme are true peptides, from different spectra, that were found in the goat reference data set from Rüther et al. (8) (C). Visual representation of the collagen-species bicluster. The cluster is zoomed in at the Pecora family level on a sample from Rangifer tarandus (RMC42); the lighter the color in the heatmap, the greater the coverage for that collagen sequence. The y-axis represents the species taxonomic tree, the x-axis represents the homology between collagen sequences, colored by relatedness. (D). Output sunburst plot of the Rangifer tarandus sample, where higher scores represent higher likelihood that the named taxon approximates the sample content. All ambiguity is retained in the final output, whether originating from the isoBLAST approach or directly from the search engine results. Overview of the ClassiCOL outcome scenarios. This schematic (created with Xmind) shows all potential outcome scenarios of the ClassiCOL algorithm in the form of a decision tree. (i) Depicts the scenario of the identification of a single species, which can be the true species or the most closely related species in the database, (ii) shows how to interpret an outcome specific to a higher taxonomic level than the species level, and (iii) visualizes how to infer a physical and/or a genetic mixture from the ClassiCOL output files. We have included “coprolite” and “bag dust” as real-life examples of physical sample mixtures analyzed with ClassiCOL.
March 16, 2025 at 9:08 AM
matthewcollins.bsky.social
4/9 ClassiCOL then resolves this ambiguity through an anti-Occam's razor approach, following the taxonomic tree and discarding branches that don't contribute unique peptides, until reaching maximum specificity. #archaeology
Demonstration of taxonomic uncertainties from a ZooMS/MS output file. (A). Unipept (36) sunburst view on Mascot result of Homo sapiens reference sample of Rüther et al. (8) showing uniqueness of peptide stretches on different taxonomic levels (Figure S1). Sunburst plot is colored via phylogenetic relatedness. (B). Histogram showing the percentage amino acid distribution between the human reference proteome (UP000005640), collected from Uniprot (red) and solely collagen sequences extracted from the same human proteome from Uniprot (blue). (C). Demonstrative overview of Homo sapiens collagen peptides, alongside the top results with similar Mascot scoring. Due to an isobaric switch the protein of origin and species of origin can change. Unfortunately, intensity prediction and the accompanying MS2PIP (63) score does not always change either, compromising this second-pass solution to further filter the peptide results.
March 16, 2025 at 9:08 AM
matthewcollins.bsky.social
4/5 🧠
Challenging peptide: COL1ɑ2 978-990 (peptide A) crucial for distinguishing reindeer from Bos/Bison.
Study shows m/z 1192.6796 is critical for Bos/Bison ID, while m/z 1166.6276 indicates reindeer even when other markers absent.
#ZooMS
Examples of peptide A (COL1ɑ2 978–990) m/z for two specimens analysed with HCl-MALDI-TOF MS. At the top, in blue, specimen PI-134 with peptide A shows 1150.6 and 1166.6 m/z peaks that are characteristic of reindeer. At the bottom, in orange, specimen PI-133 with peptide A shows 1192.7 and 1208.7 m/z peaks that are specific to Bos/Bison.
March 16, 2025 at 9:02 AM