{"id":"724c0bba-2b57-4a7d-892e-4e219cede693","arxiv_id":"2412.06550","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"Researchers report DNA fragments in crude oil and interpret them as ancient fossils, but the evidence does not rule out modern contamination.","lead":"Researchers report extracting and sequencing over three million DNA fragments from crude oil, with most matching bacteria and humans, and interpret a small fraction as ancient DNA from past species. The paper proposes oil as a new type of fossil for paleontology, but the evidence for ancient origin is not established.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No extraction blanks, no reagent controls, and a single source sample leave open that alleged ancient DNA is contamination; all historical claims depend on this.","rationale":"The reader's weakest_assumption and my own reading converge on the same load-bearing premise: the non-human, non-bacterial sequences are ancient and endogenous rather than modern contamination or bioinformatic artifacts. Everything in the paper's central claim—petroleum as a new type of fossil, DNA fossils, marine incursions, Homo erectus at 0.49 Ma, bird ancestry—depends on that premise. The evidence against it is strong and internal: no negative controls are described despite a low-biomass matrix; the only source of all 'ancient' hits is one sample while the second sample gave none; bacterial and human reads dominate in exactly the pattern expected from laboratory or reagent contamination; the absence of deamination is admitted and given an ad hoc, untested explanation; and the BLAST/E-value thresholds are applied inconsistently (Table S2 contains hits above the stated 1e-07 cutoff). The paper does include some honest caveats—'we cannot be sure whether they are oriDNA or paeDNA from other sources'—and the geochemical biomarker work appears consistent with prior regional studies, but those caveats do not rescue the strong historical conclusions. A single controlled re-extraction with blanks and an independent sample would settle whether the signal is real or contamination. The reader's verdict of REJECT is appropriate; no verdict change is needed.","tokens_in":66321,"tokens_out":2787,"duration_ms":34645,"concrete_test":"In a dedicated clean-room aDNA facility, re-extract and sequence blind: (i) at least three extraction blanks and reagent blanks; (ii) a fresh aliquot of NY201905; and (iii) an independent crude oil sample from a different well or field. Run the identical 'Mega screening method' on the blanks; if fish, bird, primate, or mollusk assignments appear in blanks at comparable frequencies, contamination is established. Also test positive samples for post-mortem damage (e.g., 5'-C-to-T deamination excess using mapDamage or DamageProfiler) and require non-human hits to survive repeat-masked, low-complexity-filtered realignment with a minimum 30-bp conserved coding segment; if these fail, the endogenous-ancient interpretation is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim requires that the rare non-human, non-bacterial BLAST hits in the pDNA dataset are endogenous ancient DNA that entered the reservoir in the geological past. The weakest step is attribution of these hits to ancient organisms rather than to contamination or artifacts. The manuscript reports no extraction blanks, no reagent controls, and no independent replication: the 402 'unique lineage' sequences and all non-indigenous hits (Tables S1–S8) come from a single sample (NY201905), while the only other sample (NY202102) yielded 'None' (Table S10). The observed profile—51% bacteria and 33% human—is the classic contamination/kitome signature for low-biomass samples. The paper's own admission that no deamination was detected ('the analysis of the pDNA sequences did not show the typical deamidation reaction often found in aDNA') directly contradicts the damage signature expected after geological timescales, and the offered explanation is untested. Several included 'aDNA' hits also fail the paper's stated E-value cutoff (e.g., 1M at E=2e-06 and 5M at E=2e-07 in Table S2), and the 'MS mode' interpretation of low-affinity hits as ancient is circular because low affinity can equally reflect unspecific short-read matches to conserved or repetitive regions. Thus the historical narratives (marine transgression, Homo erectus, bird ancestry, Pangaea dispersal) rest on unvalidated BLAST assignments.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript claims to have extracted 3,159,020 petroleum DNA (pDNA) sequences from crude oil of the Nanyang Oilfield using nanoparticle affinity bead technology, and to have classified them with a 'Mega screening method' into original in-situ DNA (oriDNA), ancient/post-depositional DNA (paeDNA), and recent environmental DNA (preDNA). From 402 sequences assigned to non-bacterial, non-human lineages, the authors infer marine transgressions, Homo erectus activity, ancient bird ancestry, and Pangaea-related dispersal, and propose petroleum as a new type of DNA fossil archive. The central positive claim is that crude oil preserves ancient DNA fragments that carry ecological and evolutionary information over geological timescales.","tokens_in":66581,"tokens_out":4803,"duration_ms":56091,"significance":"If the central claim were established, the pDNA dataset would be a genuinely novel paleontological resource and the paper would have wide impact across paleontology, petroleum geology, and evolutionary biology. The manuscript also has some legitimate strengths: the sequencing data are deposited under BioProject PRJNA1091869, the geochemical biomarker analysis follows standard petroleum geochemistry practice, and the scale of sequencing is large. However, the paper does not provide the experimental controls, damage authentication, or independent validation needed to distinguish ancient endogenous DNA from modern contamination or bioinformatic artifacts, and the historical narratives in the Results depend entirely on that distinction. As presented, the study does not support its headline claims.","major_comments":[{"comment":"The claim that non-bacterial, non-human pDNA sequences are ancient rests on a single productive sample (NY201905), and the manuscript reports no extraction blanks, reagent blanks, or negative controls. The observed composition (51% bacteria, 33% human) is the signature expected for reagent/kit contamination and laboratory or handling contamination in a low-biomass sample, and the only other sample, NY202102, yielded 'None' for non-indigenous species. Without controls, the 402 'unique lineage' sequences cannot be distinguished from modern contaminants introduced during drilling, sample transport, DNA extraction, library preparation, or sequencing, and every historical conclusion in the Results inherits this defect.","section":"Methods: DNA extraction, DNA library construction, sequencing, and NCBI nucleotide BLAST; Table S10"},{"comment":"The paper explicitly states that the pDNA sequences did not show the typical deamidation damage expected for ancient DNA, and it explains this absence with an untested hypothesis of an anaerobic, anhydrous oil-phase preservation environment. In ancient DNA research, endogeneous damage such as deamination, short fragment length distributions, and terminal overhang patterns is the primary authentication criterion; in the absence of such damage and of extraction/sequencing controls, there is no positive evidence that any recovered fragment predates the sampling date. The proposed preservation mechanism is invoked post hoc and is not independently constrained by any experiment or geochemical measurement in the manuscript.","section":"Results and Discussion: 'In our study, it was interesting to note...'"},{"comment":"The inference pipeline embeds the conclusion it is supposed to test. Step 5 of the 'Mega screening method' states that the authors 'ascertain if the sequence is oriDNA' by considering local climate, geological changes, and evolutionary principles, and the Results treat low-affinity BLAST hits as 'DNA fossils.' Low affinity is equally compatible with short reads matching conserved or repetitive regions, with divergence among unsequenced modern relatives, with sequencing error, or with spurious database matches. The manuscript offers no independent criterion that breaks this symmetry, so the classification of individual sequences as aDNA is circular rather than evidence-based.","section":"Methods: 'Mega screening method'; Results: 'aDNA with lower-affinity values...'"},{"comment":"Several marine 'aDNA' hits fail the manuscript's own stated E-value threshold of less than 1E-07. In Table S2, sequence 1M is reported at E = 2E-06 and sequence 5M at E = 2E-07, both of which are above the threshold; several entries also have affinity values below 50%. These sequences are nonetheless presented as evidence of marine transgressions. The Table S2 evidence therefore does not meet the paper's stated quality filter, and the reader is given no alignments to assess whether the matches are genuine local homologies rather than short, low-complexity, or repetitive elements.","section":"Table S2; Methods: E-value cutoff"},{"comment":"Sequences assigned to chimpanzees, gorillas, yaks, turkeys, and kiwis in a single oil sample are interpreted as ancient hominins and as descendants of Pangaea ancestors. No stratigraphic, spatial, or independent molecular evidence places any of these fragments in the geological past. Many of the primate hits in Tables S7 and S8 have near-100% identity to modern reference genomes, which is exactly what would be expected from modern contamination, and the continental-drift narrative is not testable from the reported data. These interpretations require either authenticated ancient DNA from the same samples or a completely different line of evidence.","section":"Results: Pangaea discussion; Tables S3, S7, S8"}],"minor_comments":[{"comment":"The title reads 'DNA Fragments in Crude Oil Reveals Earth's Hidden History'; the verb should agree with the plural subject 'fragments' (i.e., 'Reveal').","section":"Title"},{"comment":"The header 'Home sapiens' should read 'Homo sapiens'.","section":"Table S6"},{"comment":"The text refers to the 'cammacerane index'; this should be 'gammacerane index' as used elsewhere in the manuscript.","section":"Results and Discussion: biomarker section"},{"comment":"The accession 'OU964919.12' for sequence 10M appears malformed; accessions typically do not contain a second '.' before the version number, and this should be checked against the NCBI record.","section":"Table S2"},{"comment":"The GC column is described as '30 mm × 0.25 mm i.d.'; the first dimension is presumably 30 m, not 30 mm, and this typographical error should be corrected.","section":"Methods: Gas chromatography"},{"comment":"The paper uses the terms 'oriDNA', 'paeDNA', 'preDNA', and 'pDNA' without a single consolidated definition early in the text; defining all four terms and their intended temporal boundaries in the Introduction would improve readability.","section":"General"}],"recommendation":"reject","confidential_remarks":"The manuscript's core claims require new experiments with extraction blanks, reagent controls, independent replication, and authentic ancient-DNA damage analysis. These are not local revisions but prerequisites for the central conclusion, so rejection is appropriate for the present submission; a future study with such controls could form the basis of a new manuscript."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. This is the first published dataset of DNA extracted and sequenced from crude oil, and the reads are in BioProject; that alone makes it a resource. But the central claim — that a minority of the sequences are ancient DNA recording local geological and biological history — does not hold up on the evidence presented.\n\nWhat the paper does well: the biomarker geochemistry is standard and agrees with prior work on the Biyang Sag, and the authors are candid in several places. They say directly that no deamination was detected, that the timing of marine incursions cannot be pinned down, and that the near-absence of algal DNA contradicts the expected source of the oil. This honesty is real but does not compensate for the load-bearing problem.\n\nThe soft spot is attribution. All 402 'unique lineage' sequences used for the historical narrative come from a single sample (NY201905); the other sample yielded None. There are no extraction blanks, no reagent controls, no independent replication. A profile of 51% bacteria and 33% human is the classic signature of contamination in a low-biomass sample. Several hits in Table S2 fail the paper's own E-value cutoff (1M at 2e-06, 5M at 2e-07). The 'MS mode' threshold is author-defined, and treating low-affinity BLAST hits as ancient is circular: short reads with low identity can match conserved or repetitive regions in many unrelated taxa. The absence of deamination is not a side note; it is the expected marker of DNA surviving geological time, and the 'oil immersion' explanation is untested. Claims about yaks, turkeys, and kiwis as Pangaea descendants go far beyond what single best hits can support.\n\nWho is this for? A paleogenomicist or petroleum geochemist might want the dataset for a methodology case study or a cautionary tale. It deserves a serious referee, because the question matters and the data are public, but it should not be accepted in anything close to its current form. My verdict is reject: the historical narrative collapses without controls. A serious revision would need extraction and reagent blanks, a second independent sample, a damage analysis, and a much more constrained interpretation of the BLAST results.","headline":"First crude-oil DNA dataset, but the ancient-DNA claims don't survive contact with the methods.","tokens_in":67231,"tokens_out":2966,"would_cite":false,"duration_ms":33384,"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":"Crude oil preserves DNA fragments that can act as fossils, the paper claims, making petroleum a new archive of ancient species and events.","keywords":["petroleum DNA","ancient DNA","DNA fossils","environmental DNA","nanoparticle affinity beads","Mega screening method","Nanyang Oilfield","paleontology"],"falsifier":"Take a fresh crude oil sample from a sealed subsurface sampler so no drilling mud or surface water touches it; process it in a dedicated ancient-DNA clean room with extraction and library blanks, and sequence the drilling mud and laboratory reagents alongside it; if the fish, bird, and primate signals do not survive clean-room re-extraction, or if the same signals appear in the mud or blanks, the historical readings collapse.","tokens_in":66035,"feed_emoji":"🧬","tokens_out":10365,"duration_ms":110492,"temperature":0.7,"pith_summary":"The paper seeks to establish that crude oil is a previously untapped archive of ancient DNA: molecules that enter reservoirs over geological time can survive inside the oil phase and be read as DNA fossils. Using nanoparticle affinity bead technology on Nanyang Oilfield crude, the authors recovered 3,159,020 petroleum DNA (pDNA) sequences; most match bacteria and human DNA, but a small subset aligns to algae, fish, mollusks, birds, non-human primates, and other lineages. Because almost all original source-rock DNA (oriDNA) has been lost, the authors argue that the surviving fragments are chiefly environmental DNA that accumulated later, and that low-affinity matches therefore carry historical information about the region rather than about the oil's own source organisms. On that basis they read the data as evidence of a post-Cretaceous marine transgression, Homo erectus activity around 0.5 million years ago, an ancient bird-reptile ancestor, and molecular traces of Pangaea dispersal. If correct, petroleum reservoirs would offer a geographically widespread complement to scarce bone fossils.","feed_headline":"Crude oil yields 3.15 million DNA fragments","feed_subtitle":"Ancient fish, bird, and hominin sequences suggest oil reservoirs record Earth's hidden history.","key_machinery":"The central mechanism is the combination of nanoparticle affinity bead DNA extraction with the Mega screening method. DNA is captured directly from crude oil, repaired, sequenced, and then aligned against the full sequence database without a predefined target genome; a minimum E-value mode groups the reads into lineage subsets, and an MS mode excludes the top hit to test whether the next-best match is at least 1E-02 worse, which the authors take as evidence of a unique lineage origin. The affinity metric, Identity times Coverage converted to a percentage, then ranks each fragment: high-affinity matches are likely conserved or modern sequences, while low-affinity matches are candidates for ancient DNA. The paper also proposes an oil immersion preservation mechanism: tissue layers become encased in oil, DNA is released slowly as each layer degrades, and water extraction removes DNA from the oil-water interface, producing the observed pattern of lost original DNA and accumulated environmental DNA.","core_discovery":"The core claim is that petroleum harbors a mixture of DNA from three sources: original in situ source-rock organisms (oriDNA), ancient post-depositional environmental organisms (paeDNA), and recent or present-day organisms (preDNA), and that the ancient members of that mixture are recoverable, sequenceable, and interpretable as DNA fossils. The paper reports that the original oil-forming algae and plankton are nearly absent from the pDNA data and that the abundant signals are bacterial and human, whereas a small set of 402 unique-lineage sequences aligns to marine animals, birds, algae, and primates. The readings that carry the argument are built on sequence divergence rather than damage: 41 human mitogenome fragments diverge from modern human mitogenomes, with ID_19H placed between humans and chimpanzees and dated near 0.49 Ma, and one composite avian-reptilian fragment is proposed as possible molecular evidence of a transitional bird ancestor. The authors acknowledge that the fragments cannot by themselves be dated or tied to a stratum, and they treat the marine-transgression and Homo erectus interpretations as hypotheses that stratigraphic DNA data could strengthen or overturn.","pith_inferences":["Editorial inference: the same preservation logic would predict that tar sands, asphalt, source-rock bitumens, and reservoir formation waters also carry DNA; testing those materials would show whether petroleum DNA is a universal reservoir phenomenon or specific to the Nanyang oilfield.","Editorial inference: the paper reports no deamination damage, the usual chemical signature of ancient DNA, so the ancient label currently rests on sequence divergence rather than molecular aging; a damage-profile test would settle whether low-affinity fragments are genuinely old.","Editorial inference: the yak, turkey, kiwi, and bird-reptile assignments come from similarity searches; phylogenetic analysis with additional markers could test whether these fragments truly fall outside modern diversity.","Editorial inference: if the method holds, then oil-stained museum specimens, drill cores, and bitumen-impregnated bones may also yield ancient DNA, expanding beyond traditional fossils."],"forward_implications":["If crude oil is a DNA archive, then petroleum reservoirs become sampling sites for ancient DNA, and the 3,159,020-sequence pDNA dataset is a new paleontological resource.","If the marine-organism matches are genuine, the pDNA supports a marine transgression in the Biyang Sag after the mid-to-late Cretaceous, independent of biomarker evidence.","If the 41 divergent human mitogenome fragments are ancient, they imply that archaic Homo lineages, possibly Homo erectus, lived near the Nanyang Oilfield and that hominin mitochondrial genomes evolved faster than nuclear genomes during the ape-to-human transition.","If the yak, turkey, and kiwi matches hold, they provide molecular support for a Pangaea-era common ancestry followed by divergence after continental drift.","If the approach generalizes, a global pDNA database would connect petroleum geology, industrial genomics, and paleontology."],"supporting_citations":[{"why":"Shows that refined vegetable oil heated to 400°C can still contain amplifiable DNA, the key precedent that oil-phase liquids protect DNA.","marker":"10,11,12"},{"why":"Supplies the nanoparticle affinity bead extraction chemistry used to recover DNA directly from crude oil.","marker":"13"},{"why":"Defines the roughly one-million-year ceiling for recoverable original DNA in conventional fossils, the limit the paper aims to push past with petroleum.","marker":"9"},{"why":"Documents the expected algal and microbial source of the Biyang Sag source rocks and the marine-incursion evidence the pDNA results are compared against.","marker":"6"},{"why":"Reports the Homo erectus fossil found near Nanzhao County, the local archaeological anchor for interpreting divergent human mitogenome fragments.","marker":"19"},{"why":"Provides the ray-finned fish phylogeny used to infer that the fish-associated pDNA implies a marine transgression after the mid-to-late Cretaceous.","marker":"22"},{"why":"Defines deamidation as the standard ancient-DNA damage signature; its reported absence shapes the paper's interpretation of preservation conditions.","marker":"28"},{"why":"Supplies the ancient-DNA laboratory protocols and contamination controls the wet-lab workflow claims to follow.","marker":"29"}],"fun_headline_variants":["Crude oil hides DNA fossils from ancient fish and hominins","Oil reservoirs preserve ancient DNA, hinting at hidden history","3.15M DNA fragments in crude oil reveal ancient species","Petroleum holds DNA fossils that predate modern lineages"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the fish, bird, primate, and mollusk DNA fragments entered the oil in the geological past and survived there, rather than entering during drilling, sample handling, laboratory extraction, or sequencing, or being chance matches to short database sequences.","fun_headline_variants_meta":{"raw":{"variants":["Crude oil hides DNA fossils from ancient fish and hominins","Oil reservoirs preserve ancient DNA, hinting at hidden history","3.15M DNA fragments in crude oil reveal ancient species","Petroleum holds DNA fossils that predate modern lineages"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00019,"raw_usage":{"total_tokens":1297,"prompt_tokens":857,"completion_tokens":440,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":473,"completion_tokens_details":{"reasoning_tokens":371}},"tokens_in":473,"tokens_out":440,"duration_ms":5018,"temperature":1.0,"reasoning_tokens":371,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T19:32:37.228349+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a fresh crude oil sample from a sealed subsurface sampler so no drilling mud or surface water touches it; process it in a dedicated ancient-DNA clean room with extraction and library blanks, and sequence the drilling mud and laboratory reagents alongside it; if the fish, bird, and primate signals do not survive clean-room re-extraction, or if the same signals appear in the mud or blanks, the historical readings collapse.","supporting_citations":[],"review_version":1}