{"id":"7770e27a-d765-4bd9-b6dc-85b5d5794418","arxiv_id":"2505.08632","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"high","formal_verification":"none","parameter_count":2,"one_line_summary":"A docking paper claims a two-plant-compound combination binds BCL-2 and AKT-1 more strongly than venetoclax or melatonin, but the comparison is not apples-to-apples.","lead":"A computational docking study reports that combining two plant compounds, withaferin A and garcinol, produces more negative docking scores against cancer targets BCL-2 and AKT-1 than the single drugs venetoclax and melatonin. The authors interpret this as enhanced inhibition, but the comparison is between two molecules and one molecule, so the result is not yet evidence of better drugs.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"MLSD total scores are not commensurate with single-ligand reference docking scores, so the reported 'surpassing' values do not establish enhanced BCL-2/AKT-1 inhibition.","rationale":"The reader's weakest assumption matches the main load-bearing concern: comparing a two-ligand MLSD score to single-ligand reference scores does not establish enhanced inhibition. The paper's docking pipeline is standard, but the central claim requires that MLSD scores be interpretable as per-target affinities for the ligand pair. That premise is not defended, and the paper even provides an internal counterexample: withaferin A in the AKT-1 MLSD does not contact the melatonin pocket residues, yet the combined score is reported as a target-relevant affinity. A control experiment replacing one ligand with a decoy would directly test whether the score gain is specific. Because the reader already reached REJECT for essentially this reason, my stress-test does not change the verdict. I also note the BCL-2 PDB ID inconsistency (4WJ9 in Methods vs 8HOG in Table I), which independently undermines reproducibility, but the score-comparison problem is the more fundamental flaw.","tokens_in":7555,"tokens_out":4318,"duration_ms":45707,"concrete_test":"Run a control MLSD with the same receptor, grid box, and docking protocol, replacing garcinol with a non-interacting molecule of similar size (e.g., a solvent-exposed decoy or a similarly sized alkane). If the two-ligand total score still becomes more negative than the single-ligand venetoclax/melatonin score, then the reported 'surpassing' is an artifact of adding atoms rather than evidence of specific cooperative inhibition.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central inference from docking scores to 'enhanced inhibition' rests on comparing a two-ligand MLSD total score (-11.88 ± 0.12 kcal/mol for BCL-2, -13.74 ± 0.08 kcal/mol for AKT-1) with one-ligand scores for venetoclax (-9.73) and melatonin (-7.24). But Vina's MLSD output is a single total score over both ligands plus the receptor; it includes ligand-ligand terms and naturally grows with the number of atoms and rotatable bonds. A more negative total score does not imply tighter binding of either ligand, nor that the combined inhibition exceeds the reference drug. The paper itself reports for AKT-1 that withaferin A sits adjacent to the melatonin pocket 'without any interactions with the amino acids of the original binding site' (Section III-B), so part of the -13.74 score reflects a secondary, non-conserved contact rather than occupation of the reference site. No per-ligand affinity decomposition, no control for ligand size, and no experimental IC50 or functional assay is provided. Absent those, the abstract's and conclusion's claim of 'enhanced inhibition' is unsupported. Reproducibility is further weakened by the BCL-2 PDB ID mismatch: Methods II-A states 4WJ9, while Table I lists 8HOG.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports AutoDock Vina single-ligand and multi-ligand simultaneous docking (MLSD) of withaferin A and garcinol against BCL-2 and AKT-1, compares the MLSD binding scores with those of venetoclax and melatonin, and concludes that the combination exhibits enhanced and synergistic inhibition of both targets. The paper includes methods details, docking scores, 2D/3D interaction analyses, and a limitation statement acknowledging the need for experimental validation.","tokens_in":7836,"tokens_out":4299,"duration_ms":38435,"significance":"If the claimed enhanced dual inhibition were established, the work would offer a computational rationale for combining these two natural products in cancer therapy. The paper is clearly structured and provides reproducible command-level details, and the authors appropriately list PAINS filtering and explicit limitations. However, the central quantitative claim rests on comparing an MLSD score for two ligands with single-ligand scores of reference inhibitors, a comparison that is not physically meaningful. Because the main conclusion depends on this invalid comparison and there is no experimental or molecular dynamics support, the significance of the findings is not currently established.","major_comments":[{"comment":"The central claim that the MLSD scores (-11.88 ± 0.12 kcal/mol for BCL-2; -13.74 ± 0.08 kcal/mol for AKT-1) 'surpass' venetoclax (-9.73 ± 0.1 kcal/mol) and melatonin (-7.24 ± 0.06 kcal/mol) is based on comparing a two-ligand complex score with a one-ligand score. Vina's output for MLSD is the total energy of the receptor with both ligands and includes ligand-ligand terms; it does not measure the affinity of either individual ligand, and it naturally scales with ligand size and number of rotatable bonds. This invalid comparison is the sole quantitative evidence for 'enhanced inhibition.' The paper itself states (Section III-B) that in AKT-1, withaferin A is positioned adjacent to the melatonin pocket 'without any interactions with the amino acids of the original binding site,' so a substantial part of the AKT-1 MLSD score reflects a secondary, non-conserved contact. To support the claim, the authors would need per-ligand affinity decomposition, a size-matched control, or an experimental/functional readout.","section":"III-B; Abstract; Conclusion"},{"comment":"The BCL-2 structure is given as PDB ID 4WJ9 in Methods II-A, but Table I reports the BCL-2 PDB ID as 8HOG. These are different structures (a mouse BCL-2 in complex with a peptide versus a human BCL-2 construct). This discrepancy affects every BCL-2 result in the paper and prevents reproduction. The authors must state which structure was actually used and ensure consistency throughout.","section":"II-A; Table I"},{"comment":"The conclusion of a 'synergistic inhibitory effect' is supported only by the observation of one pi-alkyl inter-ligand interaction in the BCL-2 MLSD pose (and no inter-ligand interaction in AKT-1). A single contact is not evidence of cooperativity or synergy; no thermodynamic coupling, dose-response, or functional assay is provided. The term 'synergistic' is therefore an overinterpretation of the docking output; at most the data demonstrate simultaneous occupancy of the pocket.","section":"III-B; Discussion"},{"comment":"The manuscript includes a limitation paragraph acknowledging that computational docking requires experimental validation and that missing residues and 'speculations' may affect results. This is appropriately candid, but the abstract and conclusion still state as a finding that the combination 'reveals enhanced BCL-2 and AKT-1 inhibition.' The acknowledged limitations do not repair the invalid comparison; they underscore that the headline claim is not supported by the evidence presented.","section":"Discussion (limitations)"}],"minor_comments":[{"comment":"In the sentence 'surpassed the binding affinity of venetoclax (9.73±0.1 kcal/mol)', the minus sign is missing; it should read -9.73 ± 0.1 kcal/mol.","section":"III-B"},{"comment":"The paper reports that garcinol has one PAINS alert (Catechol A); the potential implications of this alert for the claimed target specificity are not discussed.","section":"III-A"},{"comment":"The table caption could state the number of independent docking runs used to compute the mean ± standard deviation, since this information is not provided in the Methods.","section":"Table I"},{"comment":"Figures 3 and 4 would be easier to evaluate if the shared amino acids mentioned in the captions were explicitly labeled in the rendered panels.","section":"Figures"}],"recommendation":"reject","confidential_remarks":"The PDB ID inconsistency (4WJ9 vs 8HOG) and the invalid single-vs-multi-ligand score comparison are, in my view, sufficient grounds for rejection. The manuscript's central claim is not supported even by the authors' own data, and the acknowledged limitations do not address the comparison problem. I would be open to a substantially revised manuscript that uses appropriate controls and calibrated metrics."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Seen the paper. Plain take: the docking numbers are probably real, but they don't say what the authors claim. Comparing an MLSD total score for two ligands to the single-ligand score of venetoclax or melatonin is not a measure of 'enhanced inhibition.' The total score includes both ligands and the ligand-ligand term, so it naturally gets more negative with more atoms. The AKT-1 case makes this obvious: the text says withaferin A sits next to the melatonin pocket with no interactions with the original site, so part of that -13.74 is a secondary contact.\n\nWhat's legitimately new: no one seems to have docked this specific pair against these two targets, and the paper is honest about its limitations. The workflow is standard: PubChem ligands, SwissADME PAINS filter, AutoDockTools preparation, Vina docking, visual analysis. They report means and standard deviations over runs. The limitations paragraph explicitly says experimental validation and MD are needed. That is more than some docking papers do.\n\nThe soft spots are proportionally severe. The central claim is unsupported, and there are concrete errors: the BCL-2 PDB ID is 4WJ9 in Methods but 8HOG in Table I; the venetoclax value loses its minus sign in the discussion; the abstract's 'surpassing' language is repeated in the conclusion without the caveats from the limitations. The paper calls it 'synergistic inhibitory effect' when there is no synergy measurement, no per-ligand affinity decomposition, and no experimental data. The MLSD protocol itself is established (Li and Li, 2010), so this is an application-level study with an overstated interpretation.\n\nWho is this for? A reader who wants a quick example of a two-ligand docking run might find the figures useful. But as a claim about drug candidates, it needs experimental follow-up before the 'enhanced inhibition' wording is justified. I would not cite this as evidence of anything beyond 'these ligands can be docked together.' For peer review: desk reject. The central inference is invalid on its face, and the internal inconsistencies would need fixing, but even then the conclusion would not follow. Not worth referee time.","headline":"Routine MLSD docking study whose headline claim fails: a two-ligand total score cannot be compared to single-ligand reference scores to infer enhanced inhibition.","tokens_in":8382,"tokens_out":2313,"would_cite":false,"duration_ms":22703,"reading_group":"no","serious_thinker":"yes","would_accept_peer_review":false},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Withaferin A and garcinol docked together yield more negative BCL-2 and AKT-1 binding scores than venetoclax or melatonin.","keywords":["withaferin A","garcinol","BCL-2","AKT-1","multi-ligand simultaneous docking","synergistic inhibition","cancer targets","phytochemical combination"],"falsifier":"Dock withaferin A and garcinol separately into the same BCL-2 and AKT-1 structures, add their two best single-ligand energies, and compare the sum with the MLSD score; if the sum is within the reported uncertainty, the simultaneous docking gain carries no evidence of inter-ligand synergy. A cell-based assay comparing the pair with each compound alone would then settle whether any functional inhibition matches the docking ranking.","tokens_in":7352,"feed_emoji":"💊","tokens_out":12322,"duration_ms":103833,"temperature":0.7,"pith_summary":"This paper tries to show that two plant-derived compounds, withaferin A from ashwagandha and garcinol from kokum, inhibit the cancer-survival proteins BCL-2 and AKT-1 more strongly when they are docked together than either reference inhibitor does alone. The argument is carried by multi-ligand simultaneous docking, a calculation that places both molecules in the binding pocket at once and lets them touch each other as well as the protein. The reported combined binding affinities are $-11.88 \\pm 0.12$ kcal/mol for BCL-2 versus $-9.73 \\pm 0.1$ for venetoclax, and $-13.74 \\pm 0.08$ kcal/mol for AKT-1 versus $-7.24 \\pm 0.06$ for melatonin, which the authors interpret as an additive or synergistic dual inhibition. The paper frames the result as an early in-silico step toward a two-molecule, two-target cancer strategy using widely available dietary compounds, and it calls for experimental validation.","feed_headline":"Plant compound pair beats reference drugs in cancer docking","feed_subtitle":"Simultaneous docking gives withaferin A plus garcinol more negative binding energies than venetoclax and melatonin.","key_machinery":"The load-bearing object is multi-ligand simultaneous docking (MLSD), in which two ligand molecules are docked into the same receptor pocket at the same time so that ligand-ligand contacts can contribute to the final score. In this study it is the mechanism that generates the enhanced binding affinities: the inter-ligand $\\pi$-alkyl interaction observed against BCL-2, and the side-by-side occupation of the AKT-1 pocket, are what distinguish the pair's result from two independent single-ligand dockings. The estimated binding affinity reported in kcal/mol is the quantity used to compare the pair against venetoclax and melatonin.","core_discovery":"In the paper's own terms, the central result is that multi-ligand simultaneous docking of withaferin A and garcinol produces more negative binding affinities than the commercial inhibitors: $-11.88 \\pm 0.12$ kcal/mol against BCL-2 compared with $-9.73 \\pm 0.1$ kcal/mol for venetoclax, and $-13.74 \\pm 0.08$ kcal/mol against AKT-1 compared with $-7.24 \\pm 0.06$ kcal/mol for melatonin. For BCL-2, the authors attribute the gain to an inter-ligand $\\pi$-alkyl contact between the two phytochemicals inside the same pocket used by venetoclax. For AKT-1, they report that garcinol occupies the melatonin pocket while withaferin A binds in an adjacent region, and they describe the combined effect as additive or synergistic. The authors conclude that the pair shows a synergistic inhibitory effect on both targets and treat the docking energies as evidence of dual inhibition.","pith_inferences":["A natural next calculation the paper leaves undone is to compare the MLSD score against the sum of the two single-ligand scores; that difference would separate true cooperativity from the arithmetic effect of placing two molecules in one pocket.","The same simultaneous-docking protocol could be used as a cheap screen for other natural-product pairs against oncogenic targets, generating combination candidates before any wet-lab work.","Because the authors find no inter-ligand contact in the AKT-1 pose, the AKT-1 result may reflect adjacent-pocket occupancy rather than mutual stabilization; the BCL-2 and AKT-1 mechanisms would then need different optimization strategies."],"forward_implications":["If the MLSD scores are accurate predictors, the withaferin A-garcinol pair becomes a computational lead for dual BCL-2/AKT-1 inhibition in cancers that rely on apoptosis evasion and PI3K/AKT signaling.","The reported $\\pi$-alkyl inter-ligand contact against BCL-2 implies that co-formulation or a linked hybrid of the two compounds could stabilize the bound complex better than either compound alone.","The AKT-1 pose, with garcinol in the melatonin pocket and withaferin A adjacent, suggests the pair could occupy neighboring sites without competing, which may help counter resistance caused by mutations in one binding site.","The authors' dual-inhibition conclusion points to testing the combination in cell lines that overexpress both BCL-2 and AKT-1, such as the cervical and ovarian contexts cited for the individual compounds."],"supporting_citations":[{"why":"supplies the multi-ligand simultaneous docking method that allows two ligands to occupy one pocket and interact with each other during docking.","marker":"[9]"},{"why":"provides prior evidence that withaferin A suppresses anti-apoptotic proteins including BCL-2 in cervical carcinoma cells.","marker":"[6]"},{"why":"provides prior evidence that garcinol reduces AKT phosphorylation in ovarian cancer cells.","marker":"[7]"},{"why":"documents venetoclax resistance mechanisms in BCL-2, motivating a new inhibitor search.","marker":"[5]"},{"why":"identifies PI3K/AKT pathway upregulation as a cancer mechanism, making AKT-1 a target.","marker":"[2]"},{"why":"defines the anti- and pro-apoptotic roles of BCL-2 family proteins that underlie the apoptosis rationale.","marker":"[3]"},{"why":"explains how BCL-2 blocks apoptosis by heterodimerizing with BAX and preventing cytochrome c release.","marker":"[4]"},{"why":"supplies a computational precedent for dual synergistic inhibition of two enzymes by phytochemicals.","marker":"[10]"},{"why":"gives an experimental precedent for synergistic antitumor action involving withaferin A.","marker":"[11]"}],"fun_headline_variants":["Dual plant docking outbinds venetoclax and melatonin","Withaferin A plus garcinol: stronger dual binding than reference drugs","Multi-ligand docking shows herb pair beats commercial inhibitors","Simultaneous docking: withaferin+garcinol surpass both reference drugs","Plant pair's co-docking tops BCL-2 and AKT-1 drug inhibitors"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central comparison assumes that a two-ligand simultaneous docking score can be read in the same units and with the same meaning as a single-ligand reference docking score, so that a more negative value directly indicates enhanced inhibition.","fun_headline_variants_meta":{"raw":{"variants":["Dual plant docking outbinds venetoclax and melatonin","Withaferin A plus garcinol: stronger dual binding than reference drugs","Multi-ligand docking shows herb pair beats commercial inhibitors","Simultaneous docking: withaferin+garcinol surpass both reference drugs","Plant pair's co-docking tops BCL-2 and AKT-1 drug inhibitors"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001088,"raw_usage":{"total_tokens":4622,"prompt_tokens":1096,"completion_tokens":3526,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":712,"completion_tokens_details":{"reasoning_tokens":3426}},"tokens_in":712,"tokens_out":3526,"duration_ms":26429,"temperature":1.0,"reasoning_tokens":3426,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:49:13.785270+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Dock withaferin A and garcinol separately into the same BCL-2 and AKT-1 structures, add their two best single-ligand energies, and compare the sum with the MLSD score; if the sum is within the reported uncertainty, the simultaneous docking gain carries no evidence of inter-ligand synergy. A cell-based assay comparing the pair with each compound alone would then settle whether any functional inhibition matches the docking ranking.","supporting_citations":[{"cited_title":"Withaferin A Suppresses Anti -apoptotic BCL2, Bcl-xL, XIAP and Survivin Genes in Cervical Carcinoma Cells,","cited_arxiv_id":null,"evidence_quote":"provides prior evidence that withaferin A suppresses anti-apoptotic proteins including BCL-2 in cervical carcinoma cells."},{"cited_title":"Garcinol Alone and in Combination With Cisplatin Affect Cellular Behavior and PI3K/AKT Protein Phosphorylation in Human Ovarian Cancer Cells,","cited_arxiv_id":null,"evidence_quote":"provides prior evidence that garcinol reduces AKT phosphorylation in ovarian cancer cells."},{"cited_title":"Virtual Screening of Natural Compounds as Potential PI3K -AKT1 Signaling Pathway Inhibitors and Experimental Validation,","cited_arxiv_id":null,"evidence_quote":"identifies PI3K/AKT pathway upregulation as a cancer mechanism, making AKT-1 a target."},{"cited_title":"Molecular docking studies of bioactive compounds from Annona muricata Linn as potential inhibitors for Bcl -2, Bcl -w and Mcl -1 antiapoptotic proteins,","cited_arxiv_id":null,"evidence_quote":"defines the anti- and pro-apoptotic roles of BCL-2 family proteins that underlie the apoptosis rationale."},{"cited_title":"Dual synergistic inhibition of COX and LOX by potential chemicals from Indian daily spices investigated through detailed computational studies,","cited_arxiv_id":null,"evidence_quote":"supplies a computational precedent for dual synergistic inhibition of two enzymes by phytochemicals."},{"cited_title":"Synergistic antitumor effect of 5-fluorouracil and withaferin-A induces endoplasmic reticulum stress- mediated autophagy and apoptosis in colorectal cancer cells,","cited_arxiv_id":null,"evidence_quote":"gives an experimental precedent for synergistic antitumor action involving withaferin A."}],"review_version":1}