Ancient DNA from 120-Million-Year-Old Lycoptera Fossils Reveals Evolutionary Insights

Chuan-Jia Zhu; Chun-Yan Zhou; Di Wu; Fei Liang; Gang-Qiang Cao; Jin-Yu Yang; Jun-Hui Geng; Li-Juan Zhao; Li-You Qiu; Mei-Jun Li

Ancient DNA from 120-Million-Year-Old Lycoptera Fossils Reveals Evolutionary Insights

T0 review · 5 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A fossil fish yields readable DNA after 120 million years

desk verdict A 120-million-year-old DNA claim with no authenticity controls; the filtering pipeline defines the conclusion into existence. read the letter →

arxiv 2412.06521 v1 pith:DMYBEGHO submitted 2024-12-09 q-bio.GN

Wan-Qian Zhao , Zhan-Yong Guo , Zeng-Yuan Tian , Tong-Fu Su , Gang-Qiang Cao , Zi-Xin Qi , Tian-Cang Qin , Wei Zhou

show 15 more authors

Jin-Yu Yang Ming-Jie Chen Xin-Ge Zhang Chun-Yan Zhou Chuan-Jia Zhu Meng-Fei Tang Di Wu Mei-Rong Song Yu-Qi Guo Li-You Qiu Fei Liang Mei-Jun Li Jun-Hui Geng Li-Juan Zhao Shu-Jie Zhang

This is my paper · ORCID

classification q-bio.GN

keywords ancientDNAenvironmentaloriginalinsitumegascreenmethoddeaminationtransposaseLycopteradavidimolecularpaleontology

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper claims to have recovered endogenous DNA from a 120-million-year-old fossil fish, Lycoptera davidi, from the Early Cretaceous Jehol Biota. Using a screening pipeline it calls the 'mega screen method,' it sorted 1,258,901 sequenced fragments by their best matches to known genomes and kept 243 that map uniquely and with high affinity to ray-finned fish genomes, dubbing them original in situ DNA (oriDNA). These fragments average over 100 base pairs and show no deamination, which the paper takes as evidence that closed volcanic-ash burial can preserve DNA far beyond the roughly one-million-year limit usually assumed. It also reports 10 transposase-coding sequences and proposes a genome self-renewal mechanism, 'coding region sliding replication and recombination,' to help explain rapid Cretaceous fish diversification. If the identification is right, deep-time molecular paleontology becomes possible for similarly preserved fossils.

What carries the argument

The load-bearing technique is the 'mega screen method,' a two-step bioinformatic filter. First, every sequenced fragment is searched against a comprehensive public nucleotide sequence database in 'minimum E-value mode'; an E-value below 1E-7 marks a qualified sequence, and sequences are grouped into lineage subsets such as bacteria, primates, angiosperms, and ray-finned fish. Second, 'MS mode' repeats the search with the top hit excluded; if the E-value difference between the best and second-best hits at the same taxonomic level exceeds 1E-2, the fragment is said to have a unique origin. Subset Affinity, calculated as identity times coverage, and Affinity Index thresholds of 90% and 97.5% are used to decide that the ray-finned fish subset is dominated by ancient rather than recent DNA, and geographical reasoning is then applied to call the retained sequences oriDNA.

What would settle it

Extract and sequence DNA from fresh samples of the same fossil bed in an independent laboratory that has never handled fish DNA, with extraction blanks run alongside; if the same 243 fragments do not reappear, they are contaminants. Equally decisive: search modern sediment and water from the Beipiao fossil site for these exact sequences, since finding them there would mean they are environmental DNA, not 120-million-year-old fish DNA.

Watch

Extended reading notes

Core claim

On its own terms, the paper's central discovery is that 243 DNA fragments recovered from a Lycoptera davidi fossil are original in situ DNA from the fish's genome rather than environmental contamination. The 243 sequences are classified into 180 in a 'local fish group' mostly matching cypriniform genomes, 49 matching non-local or marine fish, and 14 unplaced ray-finned fish; the paper reads the local group as a genomic connection between Lycoptera and the ancestors of carp fishes. It further reports that 10 of the fragments encode transposase and that their splicing pattern reveals a new way coding regions can be born, 'coding region sliding replication and recombination,' allowing genomes to expand without external DNA. Finally, the paper argues that the absence of deamination in these fragments shows deamination is not a necessary marker of ancient DNA, and that genomes assembled using that criterion are likely mixed genomes that should be re-evaluated.

Load-bearing premise

The argument rests on assuming that a DNA fragment whose single best match is a ray-finned fish genome comes from the 120-million-year-old fossil fish, rather than from modern contamination or from a sequence conserved across many species.

Editorial extensions

If this is right

Endogenous DNA can survive in closed volcanic-ash fossils for roughly 120 million years, far beyond the one-million-year ceiling that currently defines the field.
The 243 fragments give a direct molecular window into Lycoptera davidi and support a genomic link between this Cretaceous fish and modern carp or cypriniform genomes.
The ten transposase fragments, if correctly assigned, document a 'coding region sliding replication and recombination' mechanism by which fish genomes can generate new coding sequences without external DNA input.
Deamination should not be treated as a necessary marker of ancient DNA, so ancient-human and other genomes assembled under that criterion may contain mixed or modern sequences and warrant re-examination.
The same screening strategy could be applied to other fossils preserved in fine-grained volcanic sediment to test how general deep-time DNA preservation is.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The 243 retained fragments are probably the tip of a much larger ancient signal, because anything sharing conserved sequence with another lineage or falling below the E-value threshold is discarded, so the recovery count understates what could be present.
A decisive independent check would be to re-extract and sequence fresh fossil material in a laboratory that has never handled fish or primate DNA and to include extraction blanks; absence of the same fragments under those conditions would point to contamination rather than preservation.
If the method generalizes, the same closed-tuff preservation may allow DNA searches in other Jehol Biota fossils, such as feathered dinosaurs and early birds, where stained nuclei have already been reported.
The proposed sliding-replication mechanism makes a testable prediction: modern fish genomes should contain composite transposase genes whose junction points match the reading-frame shifts the paper illustrates.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Bottom line: the paper claims 120-million-year-old Lycoptera DNA, but the evidence doesn't come close. The sequences are selected by BLAST uniqueness to ray-finned fish, then asserted to be endogenous. No extraction blanks, no raw reads deposited, and the absence of deamination is treated as preservation when standard aDNA models predict the opposite for material that old. The conclusion is effectively built into the filtering definition.

What's actually new: the wet-lab protocol is described in detail, the sequencing effort is substantial (1.26M reads, 243 final sequences), and the 'coding region sliding replication' transposase idea is a novel speculation. The authors are transparent about the small number of sequences and the lack of a Lycoptera reference genome. That transparency is worth something.

The soft spots are more than minor. The 'mega screen method' is a similarity heuristic, not an authenticity test. A modern fish contaminant or a conserved region shared across vertebrates would pass the same filter. The geographic argument (hilly area, no lakes) doesn't rule out airborne or historical contamination, and the cofferdam portion yielded only 635 sequences and isn't used as a negative control. Claiming that no deamination is evidence of a closed system ignores that deamination is expected over 120 million years; its complete absence is a red flag. The paper goes on to argue that the Pääbo deamination criterion is wrong and that Neanderthal genomes are mixed — that's a strong claim with no support here. The panspermia paragraph at the end doesn't help.

Who this is for: a reader interested in the sociology of aDNA claims, or as a teaching example of circular filtering. The transposase mechanism might interest genome evolution folks, but only if the sequences are real.

If this came to me, I'd desk reject. There is no public raw data, no independent replication, and no negative controls. The authors should deposit the reads, include extraction blanks, and benchmark their method against modern and ancient DNA standards before asking for referee time.

Referee Report

5 major / 5 minor

Summary. The manuscript reports the extraction and sequencing of DNA from Early Cretaceous (ca. 120 Ma) Lycoptera davidi fossils and claims to have identified 243 'original in situ DNA (oriDNA)' sequences through a new bioinformatic protocol called the 'mega screen method.' The authors further report 10 transposase coding sequences and propose a novel genome mechanism ('coding region sliding replication and recombination'), as well as various evolutionary inferences about fish diversification and the origin of carp fishes. The central claim is that endogenous DNA can survive for 120 million years in fossil fish and that the identified sequences are authentic ancient DNA from Lycoptera.

Significance. If substantiated, the claim would fundamentally extend the known limits of DNA survival and would provide a new deep-time window into fish genome evolution. However, the manuscript does not meet the evidentiary standards of the ancient DNA field: the 'oriDNA' designation rests entirely on a BLAST similarity filter plus geographic plausibility arguments, with no extraction blanks, no independent replication, no damage-pattern analysis, and no deposition of raw sequencing data. The absence of deamination, which the authors treat as supportive, is instead the single most parsimonious indicator of modern contamination. The paper does provide a detailed protocol and complete sequence lists in the supplementary tables, and the authors are transparent about some limitations (e.g., the limited number of oriDNA and the inability to analyze many sequences), but these strengths do not offset the absence of authenticity criteria for the central claim.

major comments (5)

[Materials and Methods, 'The mega screen method'; Results, 'Identification of each sequence in the ray-finned fish…] The definition of oriDNA is circular. A read is classified as oriDNA if its best BLAST hit is to an actinopterygian genome and the E-value difference from the next-best hit exceeds 1E-02 (MS mode). These same reads are then reported as 'Lycoptera oriDNA' and used to infer evolutionary conclusions. A modern fish contaminant, or a conserved sequence shared across vertebrates, would pass the same filter. The thresholds (E-value < 1E-07, Affinity Index 90% and 97.5%, MS-mode difference 1E-02) are not benchmarked against any positive or negative control, so the protocol has no demonstrated ability to discriminate endogenous ancient DNA from modern environmental DNA.
[Materials and Methods, 'DNA Extraction, DNA Library Construction, and Sequencing'; Data and materials availability] No extraction blanks, library blanks, or independent replication are reported, and no raw sequencing reads are deposited. The 'cofferdam' (non-textured) portion yielded only 635 sequences and is not used as a contamination control; the paper does not compare the taxonomic composition or damage patterns of the cofferdam fraction with the fish-layer fraction. Without these controls, environmental contamination cannot be excluded. The Data and materials availability statement ('All data is available in the main text or the supplementary materials') is insufficient for a claim of this magnitude, since raw sequence data are not provided.
[Results, 'The "deamination" did not occur in Lycoptera oriDNA and some paeDNA sequences'; Results, 'The "deamination"…] The absence of deamination is used as evidence for a closed preservation system, but this is the opposite of what current models of DNA degradation predict for 120-million-year-old DNA; the absence of deamination is more parsimoniously explained by modern contamination. The manuscript's assertion that deamination is not a necessary marker of aDNA is a reasonable general caveat, but it reverses the evidentiary burden: a claim of extraordinary preservation requires positive evidence (e.g., damage gradients, fragment-length distributions, replication in independent extracts), not merely the absence of a damage signal. The further claim that Pääbo's team assembled 'mixed genomes' and that those works 'should be re-evaluated' is not supported by any data presented here and is outside the scope of this study.
[Results, 'Decoding fossil DNA: Revealing parasites and prey'] The identification of Ichthyosporea, Macrobrachium, Penaeus, and sea-spider-like sequences as oriDNA from the fish's parasites and prey rests on the absence of co-deposited animal remains in 3D X-ray scans. Absence of visible remains does not establish that these DNA fragments are endogenous; such sequences are exactly what would be expected from environmental DNA in the surrounding sediment or from contamination during handling. The inference that these are ancient dietary or parasitic associations is therefore unsupported.
[Results, 'New mechanism for generating transposase-encoding sequences'] The proposed 'coding region sliding replication and recombination' mechanism is inferred from 10 transposase-like sequences whose only link to Lycoptera is the same BLAST-based oriDNA designation. No experimental validation, phylogenetic analysis, or examination of the genomic context is provided. Because the oriDNA designation collapses, the transposase mechanism and the associated claims about Cretaceous fish diversification do not follow from the data.

minor comments (5)

[Author list and affiliations] The author list contains Chinese full-width commas (e.g., 'Gang-Qiang Cao1，Zi-Xin Qi4') that should be replaced with standard ASCII commas.
[References and notes, ref. 18] Reference 18 cites 'Lindalh, T.' but the correct spelling is Lindahl, T.; please verify all reference spellings.
[Introduction, 'deamination'] The term 'deamination' is placed in quotation marks throughout the manuscript; it is a standard chemical term and should not be quoted unless the authors intend a nonstandard meaning.
[Figure 2] The four panels of Figure 2 are difficult to interpret because the labels A-D are not explained in the legend with sufficient detail; please expand the legend to describe what is shown in each panel.
[Table S4] Table S4 is referenced in the text as containing the geographic classification, but the table itself is not included in the main text; please ensure all supplementary tables are complete and accessible.

Circularity Check

2 steps flagged · score 8.0 of 10

The 243 'oriDNA' sequences are the output of the paper's own BLAST-uniqueness and geographic-exclusion filters, then reported as endogenous Lycoptera DNA; the central claim reduces to the filter by construction.

self definitional [Materials and Methods, 'The mega screen method'; Results, 'Identification of each sequence in the ray-finned fish subset']
"If the E-value difference between two search results at the same taxonomic level (species, genus, family, order, and class) exceeds 1E-02, it suggests that the test sequence (query) belongs to the species identified in the first search result, indicating a unique origin. ... As a result, we confirmed that 243 DNA sequences were specifically mapped to the ray-finned fish genome. Consequently, we concluded that the only genealogical origin of these sequences was the ray-finned fish genome."

The MS mode operationally defines 'unique origin' as a BLAST hit whose E-value is separated from the next hit by >1E-02 at the same taxonomic level. The paper then reports the 243 sequences that pass this filter as 'specifically mapped to the ray-finned fish genome' and concludes that their 'only genealogical origin' is ray-finned fish; elsewhere these same sequences are called 'oriDNA ... likely from the Lycoptera genome.' No independent authenticity criterion is applied: no extraction blank is used as a negative control, no authentic ancient-DNA damage pattern is required, and the only reported aDNA feature, absence of deamination, is observed after the extract was repaired with PreCR Mix.
self definitional [Results, 'The Lycoptera oriDNA']
"The fossil site is currently a hilly area without lakes and rivers. These conditions indicate that most of the sequences in the local fish group cannot be preDNA. At the same time, considering freshwater fish eventually become prey to other species, it is difficult for their DNA to enter the environment and become eDNA. Therefore, we can roughly infer that most of the 180 sequences in the local fish group and the 14 sequences in the non-order group could be oriDNA."

The category 'oriDNA' is assigned to fish-like BLAST hits by invoking the untested prior that freshwater fish DNA cannot persist as eDNA in a hilly area. That prior is not a measurement; it is the classification rule. Presenting the resulting assignment as support for endogenous 120-million-year-old Lycoptera DNA re-imports the assumption as if it were an empirical finding. The 'rough inference' is definitionally equivalent to the geographic-exclusion rule used to build the oriDNA set, so the conclusion is not independently derived.

full rationale

The paper's central derivation chain is: BLAST every read against NCBI, keep reads with a unique best hit to ray-finned fish, further exclude presumed modern contamination by geographic reasoning, and call the survivors 'oriDNA likely from the Lycoptera genome.' This is a self-definitional reduction: the defining criteria of the filter (unique Actinopterygii BLAST hit and geographic plausibility) are exactly the properties later reported as the discovery. No external standard validates the filter: there are no raw-data deposits, no extraction blanks used as contamination controls, no positive/negative control for the MS mode, and the absence of deamination is reported after the DNA was treated with PreCR repair mix. The transposase mechanism, the fish-evolution conclusions, and the Panspermia speculation are all downstream of the unvalidated oriDNA label, so they inherit the circularity rather than providing independent support. The paper does not rely on a self-citation chain or an imported uniqueness theorem, so this is not a self-citation circularity; it is a definition-by-filter circularity. Because the headline count of 243 oriDNA sequences is forced by the filter's own definition, the central claim reduces to its input by construction, warranting a score of 8.

Assumptions & free parameters 3 free parameters · 5 assumptions · 2 invented entities

The paper's central inference depends on a set of unvalidated thresholds and assumptions. The thresholds (E-value, Affinity Index, MS-mode separation) are chosen by the authors and are not benchmarked against known ancient or modern DNA. The classification of sequences as endogenous rests on the assumption that unique BLAST similarity to fish excludes modern contamination, which is not established. The proposed new mechanism is inferred from the same selected sequences, adding a further layer of speculation.

free parameters (3)

BLAST E-value threshold = 1E-07
Used to define qualified sequences (QS). Chosen by the authors without benchmarking against known ancient or modern DNA.
Affinity Index thresholds = 90% and 97.5%
Used to judge whether a sequence subset is mostly ancient or mostly modern. Set by the authors in Materials and Methods without independent calibration.
MS mode E-value difference threshold = >1E-02
Used to decide whether a read has a unique taxonomic origin. The threshold is arbitrary and not validated.

assumptions (5)

domain assumption A read with a best BLAST hit to a ray-finned fish genome and a unique hit under MS mode originates from a ray-finned fish genome.
Invoked in Results, 'Identification of each sequence in the ray-finned fish subset', to convert BLAST matches into genealogical origin.
ad hoc to paper Absence of deamination is compatible with, and even supportive of, ancient DNA and is not a necessary marker of aDNA.
The paper argues against the standard deamination criterion in Results, 'The deamination did not occur in Lycoptera oriDNA', and uses lack of deamination as evidence for preservation.
domain assumption Rapid coverage by volcanic ash in calm water sealed the fish from water, preventing deamination and preserving DNA for 120 million years.
Proposed in Results, 'The mechanism of DNA preservation in Lycoptera fossils', as the explanation for survival; no direct evidence supports the sealing effect.
ad hoc to paper Since no marine transgression occurred after the Cretaceous, fish-like sequences with marine best hits must be ancient oriDNA rather than contamination.
Used in Results, 'The Lycoptera oriDNA', to classify 49 non-local fish sequences as oriDNA.
ad hoc to paper Freshwater fish DNA rarely enters the environment as eDNA because fish are prey, so most local fish-like sequences cannot be modern contamination.
Used in Results, 'The Lycoptera oriDNA', to infer that 180 local fish sequences are oriDNA.

invented entities (2)

Original in situ DNA (oriDNA)
purpose: Label for the 243 sequences claimed to be endogenous to the Lycoptera fossil.
The paper defines oriDNA as sequences that pass its own BLAST-based mega screen; no independent chemical, spatial, or replication evidence distinguishes them from contamination.
Coding region sliding replication and recombination
purpose: Proposed mechanism by which transposase coding sequences arise through frame-shifting duplication and recombination.
Inferred from 10 transposase-like fragments in the same selected sequences; no functional assay or independent genomic evidence is provided.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Ancient DNA from 120-Million-Year-Old Lycoptera Fossils Reveals Evolutionary Insights." pith.science (2026). https://pith.science/paper/DMYBEGHO

@misc{pith2026241206521,
  author       = {Pith},
  title        = {Pith review of: Ancient DNA from 120-Million-Year-Old Lycoptera Fossils Reveals Evolutionary Insights},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DMYBEGHO}},
  note         = {Machine review of arXiv:2412.06521}
}

read the original abstract

High quality ancient DNA (aDNA) is essential for molecular paleontology. Due to DNA degradation and contamination by environmental DNA (eDNA), current research is limited to fossils less than 1 million years old. The study successfully extracted DNA from Lycoptera davidi fossils from the Early Cretaceous period, dating 120 million years ago. Using high-throughput sequencing, 1,258,901 DNA sequences were obtained. We established a rigorous protocol known as the mega screen method. Using this method, we identified 243 original in situ DNA (oriDNA) sequences, likely from the Lycoptera genome. These sequences have an average length of over 100 base pairs and show no signs of deamination. Additionally, 10 transposase coding sequences were discovered, shedding light on a unique self-renewal mechanism in the genome. This study provides valuable DNA data for understanding ancient fish evolution and advances paleontological research.

Discussion (0). Continue with ORCID to comment.

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.