{"id":"618ad10d-093f-4167-9f4d-586a6bb1ead1","arxiv_id":"2412.06521","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"The authors report 243 putatively endogenous DNA fragments from 120-million-year-old Lycoptera fossils, selected by BLAST similarity and interpreted as evidence of deep-time DNA preservation and a new transposase mechanism.","lead":"This paper claims to have recovered DNA fragments from a 120-million-year-old fish fossil, far older than any previously validated ancient DNA. The evidence is a computational screen of short sequences against modern genomes, without independent authentication.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 243 'oriDNA' sequences are authenticated only by BLAST uniqueness and geographic plausibility, not by any aDNA-specific criterion; without raw data or controls, the 120-million-year endogenous origin claim is unsupported.","rationale":"The reader's weakest assumption is that a strong, unique BLAST match to a ray-finned fish genome implies the sequence originated from the fossil fish rather than from environmental DNA or a conserved region. My stress-test identifies the same load-bearing assumption, localized in the MS-mode step and the subsequent geographic filtering. The paper provides no independent authentication: no raw sequencing data, no negative controls, no replication, and no damage patterns consistent with ancient DNA. In fact, the paper explicitly states that the 243 sequences show no deamination and then argues that deamination should not be required for aDNA identification; this inverts the standard authenticity criterion without providing a validated substitute. The proposed concrete test directly probes whether the uniqueness filter is sufficient by checking non-actinopterygian hits and assessing deamination on the actual sequences. If either check fails, the central claim is unsupported. This does not change the reader's REJECT verdict, so the recommended verdict is UNCHANGED.","tokens_in":53204,"tokens_out":4299,"duration_ms":52689,"concrete_test":"Obtain the 243 sequences from Tables S3A/S4 (and, ideally, the raw FASTQ and extraction-blank data) and rerun the exact MS-mode BLAST protocol against the full NCBI nucleotide database, recording for each query the best non-Actinopterygii hit and its E-value. If any query has a non-Actinopterygii hit within 1E-02 of its top Actinopterygii hit, the uniqueness criterion fails and the sequence is not diagnosable as fish-specific. Independently, if raw reads are available, run mapDamage/PMDtools: authentic 120-million-year-old DNA should show elevated C-to-T deamination at read ends; the reported complete absence of deamination would instead indicate modern contamination, falsifying the oriDNA interpretation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim—that 243 sequences are endogenous Lycoptera DNA from 120-million-year-old fossils—rests entirely on the 'Mega screen method' (Materials and Methods, 'The mega screen method'; Results, 'Identification of each sequence in the ray-finned fish subset'). A read is assigned to oriDNA if its best BLAST hit is to Actinopterygii and the E-value difference from the next-best hit at the same taxonomic level exceeds 1E-02. This is a similarity-based origin heuristic, not an authenticity test. A modern fish contaminant, or a conserved sequence shared across vertebrates, would pass the same filter. The paper then invokes geographic reasoning ('hilly area without lakes and rivers', 'freshwater fish ... difficult ... to become eDNA') to exclude preDNA, and the reported absence of deamination to argue for closed-system preservation. Each step is unsupported: no extraction blanks are reported, no raw reads are deposited, the non-fish 'cofferdam' portions yielded only 635 sequences and are not used as a contamination control, and the complete lack of deamination is the opposite of what current aDNA models expect for authentic 120-million-year-old DNA. Because the oriDNA designation is the foundation for the transposase 'coding region sliding replication and recombination' mechanism and all downstream evolutionary conclusions, the collapse of this inference collapses the paper's central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","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.","tokens_in":53597,"tokens_out":2731,"duration_ms":31219,"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":[{"comment":"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.","section":"Materials and Methods, 'The mega screen method'; Results, 'Identification of each sequence in the ray-finned fish…"},{"comment":"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.","section":"Materials and Methods, 'DNA Extraction, DNA Library Construction, and Sequencing'; Data and materials availability"},{"comment":"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.","section":"Results, 'The \"deamination\" did not occur in Lycoptera oriDNA and some paeDNA sequences'; Results, 'The \"deamination\"…"},{"comment":"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.","section":"Results, 'Decoding fossil DNA: Revealing parasites and prey'"},{"comment":"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.","section":"Results, 'New mechanism for generating transposase-encoding sequences'"}],"minor_comments":[{"comment":"The author list contains Chinese full-width commas (e.g., 'Gang-Qiang Cao1，Zi-Xin Qi4') that should be replaced with standard ASCII commas.","section":"Author list and affiliations"},{"comment":"Reference 18 cites 'Lindalh, T.' but the correct spelling is Lindahl, T.; please verify all reference spellings.","section":"References and notes, ref. 18"},{"comment":"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.","section":"Introduction, 'deamination'"},{"comment":"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.","section":"Figure 2"},{"comment":"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.","section":"Table S4"}],"recommendation":"reject","confidential_remarks":"The manuscript makes an extraordinary claim but does not provide the standard aDNA authenticity evidence (blanks, replication, damage patterns, raw data). The circular definition of oriDNA and the lack of controls are fundamental, not cosmetic. The paper also contains an unsupported dismissal of the Pääbo team's ancient genomes, which is likely to provoke strong reactions and is not warranted by the data presented. I see no way to fix the central claim within the scope of a revision; the paper would need a completely different experimental design and independent validation. The journal may also wish to consider whether the manuscript's scope (120-million-year-old DNA) matches the journal's typical standards for molecular biology or genomics submissions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","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.\n\nWhat'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.\n\nThe 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.\n\nWho 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.\n\nIf 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.","headline":"A 120-million-year-old DNA claim with no authenticity controls; the filtering pipeline defines the conclusion into existence.","tokens_in":54107,"tokens_out":2427,"would_cite":false,"duration_ms":26328,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A fossil fish yields readable DNA after 120 million years","keywords":["ancient DNA","environmental DNA","original in situ DNA","mega screen method","deamination","transposase","Lycoptera davidi","molecular paleontology"],"falsifier":"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.","tokens_in":53032,"feed_emoji":"🧬","tokens_out":7778,"duration_ms":77656,"temperature":0.7,"pith_summary":"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.","feed_headline":"Fossil fish yields DNA after 120 million years","feed_subtitle":"A new screening method pulls 243 endogenous fragments from a Cretaceous fossil and reopens deep-time evolution.","key_machinery":"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.","core_discovery":"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.","pith_inferences":["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."],"forward_implications":["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."],"supporting_citations":[{"why":"States the current field limit of roughly one million years for amplifiable ancient DNA, the limit this paper claims to break.","marker":"[1]"},{"why":"Provides the two-million-year-old environmental DNA example and the deamination baseline the paper argues against.","marker":"[8]"},{"why":"Supplies the nanoparticle affinity bead DNA extraction method used on the fossil powder.","marker":"[17]"},{"why":"Exemplar of the deamination-based ancient DNA criterion that the paper challenges as not necessary.","marker":"[19]"},{"why":"Assigns Lycoptera to Osteoglossiformes, the taxonomic framework used to interpret the fish's relationships.","marker":"[25]"},{"why":"Dates cypriniform genome origins to the early Jurassic, the comparison point for the 180 carp-like oriDNA fragments.","marker":"[29]"},{"why":"Establishes that transposons make up over 40% of fish genomes, motivating the transposase search.","marker":"[31]"},{"why":"Treats transposon sequences as 'protein fossils,' the interpretive frame for the fossil transposase fragments.","marker":"[33]"},{"why":"Lists the conventional mechanisms, horizontal transfer and duplication, that the proposed sliding replication mechanism is offered as an alternative to.","marker":"[34]"}],"fun_headline_variants":["120M-year-old fish fossil yields 243 DNA fragments","Cretaceous fish DNA reveals evolutionary insights after 120M years","Ancient DNA extracted from 120-million-year-old Lycoptera fossil","New method recovers 120M-year-old DNA from Lycoptera","Deep-time DNA: 120M-year-old fish fossil yields 243 fragments"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"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.","fun_headline_variants_meta":{"raw":{"variants":["120M-year-old fish fossil yields 243 DNA fragments","Cretaceous fish DNA reveals evolutionary insights after 120M years","Ancient DNA extracted from 120-million-year-old Lycoptera fossil","New method recovers 120M-year-old DNA from Lycoptera","Deep-time DNA: 120M-year-old fish fossil yields 243 fragments"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00063,"raw_usage":{"total_tokens":2882,"prompt_tokens":891,"completion_tokens":1991,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":507,"completion_tokens_details":{"reasoning_tokens":1899}},"tokens_in":507,"tokens_out":1991,"duration_ms":14863,"temperature":1.0,"reasoning_tokens":1899,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T19:33:40.549748+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"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.","supporting_citations":[],"review_version":1}