{"id":"f69aa4f3-8b11-49b6-8773-97ad7af78290","arxiv_id":"2412.06378","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":2,"one_line_summary":"A critique argues that environmental DNA contamination and unreliable deamination criteria undermine many ancient DNA results, including the Tianyuan Cave Man genome.","lead":"This paper claims that DNA extracted from fossils is frequently contamination from later environments, not the original animal's DNA, making the standard 'deamination' marker unreliable. A smart generalist should read it because it challenges the reliability of famous ancient human genomes, including the Tianyuan Cave Man genome.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The inference that the 22 'Pan-unique' Lycoptera fragments are ancient hominin eDNA rests on BLAST top-hit ranking alone; without controls, these short, low-identity, non-deaminated fragments cannot carry the claim that the Tianyuan genome is a cross-species mixture.","rationale":"The reader's weakest-assumption analysis correctly identifies the hinge: the claim that Lycoptera fragments matching Pan genomes are ancient hominin DNA is not established, and the Tianyuan conclusion depends entirely on it. My stress-test converges on the same point and sharpens it: BLAST top-hit ranking with a 1E-02 E-value gap is not sufficient to prove species origin for short, low-identity, non-deaminated fragments, especially without negative controls or mapping to the source fossil's own lineage. The paper's general cautions about eDNA infiltration and the unreliability of deamination as a universal age marker are plausible and consistent with existing literature, and the proposed 'unique match plus stringent E-value' screening rule is a reasonable methodological suggestion in principle. However, the manuscript does not validate that rule on known positive and negative controls, and its central empirical demonstration consists of self-described fragments from a separate unreleased preprint. Because the strongest claim is stated as a likely finding about the Tianyuan genome and a sweeping call for reevaluation, the evidence falls short. The reader's REJECT verdict is appropriate; I recommend no change.","tokens_in":6494,"tokens_out":2534,"duration_ms":33239,"concrete_test":"Re-run the MS-mode screen on the 22 'Pan-unique' fragments from Table S3B of reference 11 after adding two controls: (1) length- and GC-matched random/shuffled sequences, and (2) a Lycoptera or other fish genome as an additional BLAST database entry. If any random control is called 'Pan-unique' under the same E-value thresholds, or if any real fragment maps to fish DNA with an E-value equal to or better than its Pan hit, the uniqueness criterion has no discriminative power and the hominin-origin inference fails.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim on pages 3-4 is that DNA fragments from 120-million-year-old Lycoptera fossils that BLAST-match Pan genomes must be ancient Homininae eDNA, and this is used to conclude that the Tianyuan Cave Man genome is likely a mix of genomes from other genera and that nearly all studies using Pääbo's method need reevaluation. The load-bearing step is the assertion that the 22 'uniquely Pan' fragments are hominin in origin. This is not demonstrated. The fragments are short, can have Identity or Cover below 90%, and show no detectable deamination. BLAST E-values on such fragments are not a reliable taxonomic assignment: short sequences can top-hit a particular genome because of conserved elements, low-complexity regions, or alignment artifacts, and a 1E-02 E-value gap to the second hit does not establish biological origin. The paper provides no negative controls, no mapping of these fragments back to Lycoptera or fish genomes, and no demonstration that the fragments carry hominin-specific substitutions. The absence of deamination actually argues against ancient age unless a preservation mechanism is shown; the volcanic-tuff claim is asserted for the fossils but not shown for these specific fragments. If the fragments are instead endogenous fish sequence, microbial contamination, or alignment artifacts, the entire contamination conclusion for Tianyuan collapses.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper is a critical methodology review of ancient DNA (aDNA) research following the approach of Pääbo's team. It argues that fossil bones are open systems into which environmental DNA (eDNA) continuously infiltrates, that probe enrichment and BLAST-based screening are insufficient to exclude eDNA, and that deamination is not a reliable criterion for identifying original in situ DNA. The central empirical claim, developed in the section 'Fossils Contain Many DNA Fragments from Ex Situ Species,' is that DNA fragments from 120-million-year-old Lycoptera fossils that BLAST-match Pan genomes, while lacking deamination, must originate from ancient Homininae and therefore that the assembled 'Tianyuan Cave Man' genome is likely a mixture of genomes from different genera and periods, with the findings of nearly all studies using Pääbo's method requiring reevaluation. The paper also proposes stricter BLAST criteria, including an E-value below 1E-07 and a 1E-02 gap to the second hit, for species-unique assignment.","tokens_in":6766,"tokens_out":3949,"duration_ms":44796,"significance":"If the central empirical claims were established, the paper would have considerable significance for molecular paleontology, because it would imply widespread contamination in published ancient human genomes. The paper does identify legitimate concerns: fossils are porous, eDNA can enter bone over geological time, deamination is not a universal or sufficient marker of ancient origin, and short DNA fragments are intrinsically difficult to assign taxonomically. However, the manuscript's own load-bearing evidence is not demonstrated: the key Lycoptera fragment analyses are confined to the authors' unpublished preprints (refs 11 and 12), no negative controls or quantitative validations are presented, and the inference from BLAST top-hit rankings to 'ancient Homininae eDNA' is unsupported. The paper also contains no reproducible code or accessible data beyond a request-based offer, and its proposed thresholds are introduced without calibration. The deamination critique is partially grounded in Lindahl's work and deserves consideration, but the broad claim that nearly all Pääbo-method studies require reevaluation is disproportionate to the evidence supplied.","major_comments":[{"comment":"The identification of the 22 'Pan-unique' Lycoptera fragments as ancient hominin eDNA is not established. The fragments are short, can have Identity or Cover below 90%, and show no detectable deamination, yet the paper concludes that they 'must originate from ancient Homo species or close relatives' solely from BLAST top-hit ranking with an E-value gap of 1E-02 to the second hit. Short queries can top-hit a particular genome through conserved sequences, low-complexity regions, or alignment artifacts, and BLAST E-values do not by themselves determine biological origin. The manuscript provides no negative controls, no mapping of the fragments back to Lycoptera or fish genomes, no demonstration of hominin-specific substitutions, and no assessment of whether the fragments could be alignment artifacts. Since this inference is the sole empirical basis for the Tianyuan-mixture conclusion, the central claim collapses without it.","section":"§3 (Fossils Contain Many DNA Fragments from Ex Situ Species), pp. 3-4 and Figure 1"},{"comment":"The main empirical evidence for the contamination mechanism is contained in the authors' own unpublished preprints (refs 11 and 12), not in this manuscript. The manuscript states that the 22 fragments are listed in Table S3B of reference 11 and that full sequencing data are 'available upon request,' but no data, code, or stable supplementary material is provided here. This makes the central claim unverifiable and effectively circular: the criteria used to define 'Pan-unique' fragments are established only in those preprints and are then cited as though they validate the present conclusions. The absence of accessible data is a load-bearing reproducibility problem, not a stylistic one.","section":"§3, references 11 and 12, Table S3B of reference 11"},{"comment":"The absence of detectable deamination cannot serve as evidence that the Lycoptera fragments are ancient Homininae DNA. Modern contamination also lacks deamination, and the paper does not demonstrate a preservation mechanism that would prevent deamination specifically in these fragments. The volcanic-tuff description is asserted for the Lycoptera fossils in general and is not tied to the sequenced fragments. Consequently, the logical chain 'no deamination + not 100% identical to Pan = early Homininae origin' is invalid: sequencing error, missing reference diversity, conserved non-hominin sequence, or recent contamination are equally consistent with the observations.","section":"§3, p. 4 and §5 (\"Deamination\" Should Not Be Viewed as an Essential Criterion), pp. 5-6"},{"comment":"The 15-base sequence CCTATAGCACCCCCTC is used to argue that the 'Tianyuan Cave Man' mitogenome may be mixed with foreign fragments, but a 15-mer that matches many taxa is expected by chance alone, and the figure does not show that such matches occur among the actual sequencing reads or that they influence the assembled mitogenome KC417443. No quantitative comparison (e.g., the number of reads matching non-target taxa, or a test of whether the observed 15-mer is overrepresented in the library) is provided. This evidence therefore cannot support the conclusion that the mitogenome is contaminated.","section":"§4 (Probe Enrichment Does Not Exclude Sequences from Ex Situ Species), p. 5 and Figure 2"},{"comment":"The proposed BLAST criteria are ad hoc and uncalibrated. The manuscript recommends an E-value below 1E-07 and an E-value difference greater than 1E-02 between the top hit and the second hit, but it provides no derivation, no error model, and no validation on datasets of known composition. These thresholds also do not resolve the central difficulty, because reference databases are incomplete and short fragments can satisfy the criteria while still being non-diagnostic. As presented, the proposed method is a set of rules rather than a tested analytical framework.","section":"§7 (Exploring aDNA Research Methods), p. 7"}],"minor_comments":[{"comment":"The author name is misspelled as 'Lindalh' in reference 14; it should be 'Lindahl'.","section":"References, ref. 14"},{"comment":"Reference titles are truncated (e.g., 'The deep population history of northern...' and 'Ancient genomes revealed the complex hu...'), and citation formatting is inconsistent throughout.","section":"References, refs. 3 and 4"},{"comment":"The terms 'MS mode,' 'minimal E-value mode,' and 'MS Model' are used without definition in this manuscript; they should be defined or a precise pointer to the method in reference 11 should be given.","section":"Throughout, pp. 3-5"},{"comment":"Informal passages such as 'As one observes the rain this morning nourishing the Cretaceous rocks' and the phrase 'less old and more new molecules' are not appropriate for a journal article and should be rewritten.","section":"Introduction, p. 3"},{"comment":"The statement that artificially synthesizing and faking the fragments could cost millions of dollars is irrelevant to the scientific argument and should be removed.","section":"§3, p. 4"},{"comment":"The caption refers to sequence IDs (379-533, 290-378, 534-555) but the manuscript does not provide a table mapping these IDs to sequences; readers should be referred to a stable supplementary file rather than a preprint.","section":"Figure 1 caption"}],"recommendation":"reject","confidential_remarks":"The manuscript's most consequential empirical claims depend entirely on the authors' own unpublished arXiv preprints (refs 11 and 12), with data 'available upon request.' This is a disclosure and reproducibility concern that the editor may wish to weigh beyond the technical assessment. The paper also sits awkwardly with the journal's scope: it is framed as a methodological critique, but its central assertion is a re-interpretation of a specific published ancient genome. The deamination discussion has some independent grounding, but the paper's headline conclusion that nearly all Pääbo-method studies require reevaluation is not supported by the evidence presented."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read this if you want a concise restatement of two legitimate worries in ancient DNA work: environmental DNA can infiltrate fossils, and deamination is not a universal proof of age. The authors are right to recommend broader BLAST searches and stricter E-value thresholds; that is the useful kernel.\n\nThe trouble is the paper's central empirical claim. On pages 3–4 they assert that 22 short BLAST hits from 120-million-year-old Lycoptera fossils are ancient Homininae eDNA, and from that they conclude the Tianyuan Cave Man genome is likely a mixture of genomes from other genera. That leap fails. The hits have identity or cover below 90%, no negative controls are shown, and no attempt is made to map the fragments back to fish or microbial genomes. BLAST top-hit ranking on fragments this short is not a taxonomic assignment; conserved elements and low-complexity regions can produce exactly this pattern. The absence of deamination is not evidence for ancient origin—under the standard model it argues the opposite—and the 'volcanic tuff' preservation story is asserted for the fossils but never tied to these specific fragments. The single 15-bp sequence matching many species is a nice illustration of short-sequence multi-matching, but it does not show that any published ancient genome is contaminated.\n\nThe paper's reliance on its own unpublished preprints (refs 11 and 12) and on 'data available upon request' is a serious problem. The cost-of-synthesis argument is a red herring. None of this is to say the authors are dishonest; they may believe their data, but they have not made a falsifiable case.\n\nThe first part of the paper could serve as background reading for students, and the method suggestion is worth a footnote. But as a research article the evidence is not self-contained. I would send it to peer review only if the authors provide their full dataset and appropriate controls; otherwise the central claim is unsupported. In my next twelve months I would not cite this paper for any empirical fact.","headline":"A cautionary commentary with a credible general point, but its headline claim about Tianyuan rests on unpublished BLAST hits and a misreading of deamination.","tokens_in":7334,"tokens_out":4086,"would_cite":false,"duration_ms":41742,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper argues that standard ancient-DNA identification—built on probe enrichment, target-genome BLAST, and deamination filtering—can mistake environmental DNA for the fossil's own DNA, and that the Tianyuan Cave Man genome is therefore…","keywords":["ancient DNA","environmental DNA contamination","deamination","BLAST E-value","Tianyuan Cave Man","Lycoptera fossils","in situ DNA","fossil taphonomy"],"falsifier":"Re-run the Lycoptera data against an updated all-genome database and check whether the 22 Pan-unique fragments still top-hit Pan with an E-value gap greater than 1e-02; then attempt independent retrieval of the same fragments from fresh Lycoptera powder in a dedicated clean lab, and test whether identical fragments also appear in surrounding sediments, groundwater, or modern reagents. If the fragments do not reproduce or also appear in modern water, the ancient-hominin interpretation fails.","tokens_in":6250,"feed_emoji":"🧬","tokens_out":4168,"duration_ms":44446,"temperature":0.7,"pith_summary":"This review challenges the assumption that DNA recovered from a fossil is the original DNA of the single individual to which the fossil belongs. The authors argue that fossils are porous rock bodies continuously flushed by water, so environmental DNA from later periods can repeatedly enter and replace the original DNA. They use their own screens of 120-million-year-old Lycoptera fish fossils to claim that fragments matching ape and chimpanzee genomes, with no detectable deamination damage, are ancient hominin DNA that infiltrated the rock. From this they conclude that the assembled genome of the so-called Tianyuan Cave Man is likely a mixture of unclassified Homo species and modern humans, and that nearly all studies built on the standard method deserve reevaluation. If correct, many published ancient-human genome interpretations would need to be revisited.","feed_headline":"Fossil DNA may be mostly later contamination","feed_subtitle":"A critique argues that water flushes foreign DNA into fossils, so deamination-based filters mislabel contaminants as ancient.","key_machinery":"The argument is carried by two instruments: a sequence-uniqueness screen using BLAST against all known genomes, requiring the target species to be the top hit with an E-value below 1e-07 and an E-value gap of more than 1e-02 over the second-best match, and a physical model of fossils as porous rocks that continuously exchange old and new environmental DNA through wet-dry cycles, named 'less old and more new molecules'. The authors apply this screen to Lycoptera fossil data and interpret the resulting Hominoidea- and Pan-matching fragments as ancient hominin DNA, which they then use to argue that the standard method's reliance on deamination as an authenticity marker is misplaced.","core_discovery":"The central claim is that fossil bones behave as open, water-flushed rock: environmental DNA continuously enters internal cavities after burial, so the DNA recovered from a fossil is not necessarily the original DNA of the fossil's species. On this basis the authors argue that the widely used ancient-DNA workflow mislabels foreign fragments as endogenous: probe capture at 62–65°C is not species-specific for fragments longer than about 30 base pairs; BLAST searches restricted to the target species' genome admit environmental lookalikes; and deamination filters select damaged fragments with high E-values while excluding well-preserved original DNA. The authors report that Lycoptera fossils yielded more than 230,000 fragments hitting Hominoidea genomes and 22 fragments uniquely matching the Pan genome, none with detectable deamination, and they take these as evidence that hominin DNA has been seeping into fossils since the fossils formed. The conclusion follows: the so-called Tianyuan Cave Man genome is likely a mix of DNA from multiple Homo species, modern humans, or even other genera, and the findings of studies using the same method should be reevaluated.","pith_inferences":["A direct test of the contamination hypothesis would be to re-analyze the Tianyuan and other published ancient-DNA datasets with the proposed all-genome BLAST screen: if most assigned fragments fail the uniqueness criteria, the paper's central worry is confirmed; if they remain uniquely assigned, the worry is at least bounded.","The permeability argument implies that even perfectly clean laboratory conditions cannot certify in-situ origin, shifting the burden of proof to depositional context and taphonomic sealing.","If deamination is not a necessary criterion, some claims of extremely old DNA may be under-estimated rather than over-estimated, because well-preserved sealed fossils would be excluded by the standard filter.","The same 'less old and more new molecules' mechanism suggests that sediment, rock voids, and petroleum reservoirs near fossil sites are promising sources of ancient DNA, extending the paper's proposal beyond fossils proper."],"forward_implications":["If the central claim is right, published ancient-human genomes assembled with the standard capture-and-deamination workflow cannot be treated as representing single individuals.","Ancient-DNA reference databases should be expanded from the target species to all known genomes before any fragment is assigned to a species.","Fragments that fail the proposed uniqueness criteria, including those with E-value gaps smaller than 1e-02, should be discarded before genome assembly.","Deamination should not be used as a positive marker of ancient origin, and heavily deaminated fragments should be avoided because they produce inflated E-values and unreliable BLAST hits.","The Tianyuan Cave Man mitogenome and chromosome 21 assembly are specifically called into question as potentially mixed with DNA from many other species."],"supporting_citations":[{"why":"Supplies the Tianyuan Cave Man genome that is the central target of the paper's reassessment, including the short fragment-length statistics and the single example sequence the authors re-BLAST against all genomes.","marker":"[1]"},{"why":"Supplies the cave bear example in which significant human mitochondrial DNA was found in an animal fossil, motivating the cross-species contamination problem.","marker":"[2]"},{"why":"Supplies the Lycoptera fossil dataset, the internal-space volume calculation, the less-old-more-new model, and Table S3B containing the Pan-matching fragments with no detectable deamination.","marker":"[11]"},{"why":"Supplies the MS model for uniquely matched fragments and the cost argument used to rule out deliberate fabrication of the matching sequences.","marker":"[12]"},{"why":"Supplies the probe-enrichment protocol with hybridization temperatures of 62–65°C, which the paper argues is not species-specific for longer fragments.","marker":"[13]"},{"why":"Supplies the chemical principle that deamination requires liquid water, used to argue that deamination is not a uniform or necessary ancient-DNA marker.","marker":"[14]"},{"why":"Supplies the sedimentary ancient-DNA approach that the paper extends toward a broader search for DNA in rocks and petroleum.","marker":"[18]"}],"fun_headline_variants":["Fossil DNA often isn't from the fossil","Deamination filters mislabel contaminants as ancient","Bones soak up foreign DNA, skewing ancient studies","Water flushes foreign DNA into fossil bones"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The inference that fish-fossil DNA fragments matching ape and chimpanzee genomes are ancient hominin DNA depends on assuming that BLAST top-hit ranking with the stated E-value criteria reflects biological origin, that absence of detectable deamination rules out modern contamination, and that the fragments are not conserved sequences shared across taxa or alignment artifacts.","fun_headline_variants_meta":{"raw":{"variants":["Fossil DNA often isn't from the fossil","Deamination filters mislabel contaminants as ancient","Bones soak up foreign DNA, skewing ancient studies","Water flushes foreign DNA into fossil bones"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000167,"raw_usage":{"total_tokens":1278,"prompt_tokens":984,"completion_tokens":294,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":600,"completion_tokens_details":{"reasoning_tokens":234}},"tokens_in":600,"tokens_out":294,"duration_ms":3661,"temperature":1.0,"reasoning_tokens":234,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T19:43:12.110572+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the Lycoptera data against an updated all-genome database and check whether the 22 Pan-unique fragments still top-hit Pan with an E-value gap greater than 1e-02; then attempt independent retrieval of the same fragments from fresh Lycoptera powder in a dedicated clean lab, and test whether identical fragments also appear in surrounding sediments, groundwater, or modern reagents. If the fragments do not reproduce or also appear in modern water, the ancient-hominin interpretation fails.","supporting_citations":[{"cited_title":"Fu et al., DNA analysis of an early modern human from Tianyuan Cave, China","cited_arxiv_id":null,"evidence_quote":"Supplies the Tianyuan Cave Man genome that is the central target of the paper's reassessment, including the short fragment-length statistics and the single example sequence the authors re-BLAST against all genomes."},{"cited_title":"Dabney et al., Complete mitochondrial genome sequence of a Middle Pleistocene cave bear reconstructed from ultrashort DNA fragments","cited_arxiv_id":null,"evidence_quote":"Supplies the cave bear example in which significant human mitochondrial DNA was found in an animal fossil, motivating the cross-species contamination problem."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the Lycoptera fossil dataset, the internal-space volume calculation, the less-old-more-new model, and Table S3B containing the Pan-matching fragments with no detectable deamination."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the MS model for uniquely matched fragments and the cost argument used to rule out deliberate fabrication of the matching sequences."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the probe-enrichment protocol with hybridization temperatures of 62–65°C, which the paper argues is not species-specific for longer fragments."},{"cited_title":"Lindalh, Instability and decay of the primary structure of DNA","cited_arxiv_id":null,"evidence_quote":"Supplies the chemical principle that deamination requires liquid water, used to argue that deamination is not a uniform or necessary ancient-DNA marker."},{"cited_title":"Zhang et al., Denisovan DNA in Late Pleistocene sediments from Baishiya Karst Cave on the Tibetan Plateau Science 370, 584-587 (2020)","cited_arxiv_id":null,"evidence_quote":"Supplies the sedimentary ancient-DNA approach that the paper extends toward a broader search for DNA in rocks and petroleum."}],"review_version":1}