{"id":"966f0386-63ad-4814-b0f3-8016bcb67a70","arxiv_id":"2508.06720","paper_version":3,"verdict":"UNVERDICTED","confidence":"LOW","novelty_score":5.0,"correctness_risk":"unknown","formal_verification":"none","parameter_count":0,"one_line_summary":"MT-REXEE with a new chain-growth swapping approach yields faster-converging relative binding free energy estimates for a flexible FabB/ACP complex with two alternate binding pockets.","lead":"This paper applies an enhanced sampling method, multiple topology replica exchange of expanded ensemble (MT-REXEE), to compute binding free energy differences for flexible acyl chains that can bind in two pockets of the FabB enzyme. The authors report faster convergence of relative binding affinity estimates, achieved without predefining the binding pockets.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Faster convergence does not prove unbiasedness: the new swapping move needs a validity check against a known reference.","rationale":"The reader's verdict is UNVERDICTED due to abstract-only review, and our concern does not change that: the available evidence is insufficient to rule out bias in the new swapping move. We partially agree with the reader's weakest assumption: the chain-length coordinate must connect the two pockets for the method to work, but we identify a more fundamental prerequisite—the swapping move itself must be unbiased. Even if the chain-length route is valid, a biased swap produces fast convergence to incorrect free energies, which would negate the central claim. The concrete test we propose is a standard validation step for enhanced sampling methods: check the sampled distribution against the known target and verify detailed balance. This would settle the concern if full text or code were available. Since neither is available, the verdict remains UNVERDICTED, but the concern should be flagged for any full-text review or reproducibility assessment.","tokens_in":811,"tokens_out":5130,"duration_ms":68882,"concrete_test":"Run the open-source implementation (if provided) on a simple two-state model with a known free energy difference, or on the FabB/ACP system at a single chain length where the pocket affinity can be computed by long unbiased simulations. Compare the empirical distribution of the alchemical coordinate from MT-REXEE with the target distribution defined by the expanded-ensemble weights (e.g., via a chi-square test on the visited chain-length histogram). Also verify that the swap acceptance ratio matches the Metropolis-Hastings ratio for the proposed move. If the empirical distribution deviates beyond statistical error or the acceptance ratio is inconsistent, the new swapping move is biased; if both pass, the unbiasedness concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that the new swapping approach in MT-REXEE yields significantly enhanced sampling, evidenced by faster convergence of free energy estimates for kinetically separated binding pockets. For that claim to be meaningful, the free energy estimates must converge to the correct values. The abstract provides no statement that the new swapping move satisfies detailed balance or preserves the target expanded-ensemble distribution. If the swapping scheme introduces a systematic bias (e.g., through an invalid acceptance criterion, an adaptive weight update that is not equilibrated, or a non-terminating replica-exchange proposal), the estimates could converge rapidly to incorrect free energies, making the observed speed-up an artifact. This is load-bearing because the novel contribution of the paper is precisely the swapping protocol; its only reported validation is a speed metric, which cannot distinguish fast convergence to the right answer from fast convergence to the wrong answer. The reader's identified assumption—that the chain-length coordinate visits transitions between pockets—is secondary: even if that mechanism operates, an invalid swap invalidates the result. Without access to the full text, no acceptance-ratio check or independent cross-validation is available.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript describes an application and implementation of MT-REXEE (multiple-topology replica exchange of expanded ensemble) to compute relative binding free energies between two alternative binding pockets of the FabB/ACP complex. The central claim is that a new swapping approach within MT-REXEE yields significantly enhanced sampling, demonstrated by faster convergence of free energy estimates for kinetically separated binding pockets. The abstract further claims that the method achieves unbiased sampling of alternate configurational states without prior pocket definitions by letting each simulation visit chain lengths where transitions between pockets occur. The abstract contains only a qualitative statement of enhanced sampling; no quantitative results, error bars, or convergence criteria are provided.","tokens_in":1032,"tokens_out":3127,"duration_ms":40722,"significance":"If the claimed behavior is correct, this is a valuable methodological advance: the approach would address a recognized bottleneck in alchemical free energy calculations for flexible complexes, avoid hand-crafted collective variables, and provide an open-source implementation. These are real strengths. However, as presented, the central claim rests on a speed metric only, with no demonstration that the accelerated estimates are unbiased. The significance of the contribution cannot be assessed from the abstract alone; a quantitative comparison with an independent reference, or at least a statement of a rigorous correctness check, is required.","major_comments":[{"comment":"The core claim, 'significantly enhanced sampling ... demonstrated by faster convergence of free energy estimates,' is stated without quantitative support. No numerical free energy differences, error bars, replicate analyses, or convergence thresholds are reported. A reader cannot distinguish fast convergence to the correct value from fast convergence to a biased value. The abstract should report a specific convergence metric (e.g., time to a given uncertainty, agreement with an independent calculation, or variance across repeats) for both the new and baseline approaches.","section":"Abstract"},{"comment":"Faster convergence does not by itself establish unbiasedness. The novel swapping move must be shown to preserve the target expanded-ensemble distribution, for example by proving detailed balance or by comparing the final free energy estimates to an independent reference calculation. If the acceptance criterion is approximate, or if the adaptive weight updates are not equilibrated, the observed speed-up could be an artifact of systematic bias. The abstract currently offers no statement addressing this load-bearing correctness concern.","section":"Abstract"},{"comment":"The mechanistic premise that 'allowing each simulation to visit chain lengths where transitions between the pockets occur' is presented as the reason for the enhanced sampling. This is an assumption about the system's free-energy landscape; it should be supported by simulation data, e.g., observed pocket-occupancy transitions as a function of chain length. Without such evidence, the explanation for the claimed improvement is not established.","section":"Abstract"}],"minor_comments":[{"comment":"The phrase 'unbiased sampling' is a strong claim. Consider qualifying it as 'unbiased in the tested cases' or adding a reference to a rigorous test, since no statistical or formal evidence is visible in the abstract.","section":"Abstract"},{"comment":"Some terms, such as 'MT-REXEE' and 'the new swapping approach,' are defined only by reference to prior work. A brief operational definition in the abstract would help readers assess the novelty and scope of the method.","section":"Abstract"}],"recommendation":"uncertain","confidential_remarks":"This is an abstract-only review, so I cannot reach a definitive accept/reject decision. If the full manuscript already contains a detailed-balance proof or an independent reference calculation, the major concerns above may be addressed, but the abstract should still include at least one quantitative convergence metric. I recommend sending the full text for assessment before a final editorial decision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Honest take: this is a plausible extension of MT-REXEE with a genuinely useful case study, but the abstract alone cannot support the central claim of unbiased faster convergence. The load-bearing question is whether the new swapping move preserves the target distribution.\n\nWhat's new: a swapping approach for MT-REXEE applied to alternate pockets in FabB/ACP without predefined collective variables. That's a legitimate niche—flexible complexes with high barriers are hard for alchemical methods. The open-source implementation is a plus, and the methodological story is coherent.\n\nWhere I have concerns: the only reported validation is faster convergence of relative binding free energies. Faster convergence to the wrong answer is a real risk for any new replica-exchange move. The stress-test note is right: an invalid acceptance criterion or a non-equilibrated adaptive weight update could produce exactly this artifact. The abstract gives no indication that detailed balance is preserved or that results were cross-checked against a known reference. That's a load-bearing soft spot, not a minor quibble. The secondary assumption about chain length visiting transitions between pockets is less worrying; even if true, it doesn't fix an invalid swap.\n\nI want to be fair: this is an abstract-only review. The full manuscript may well include a detailed balance check, acceptance-ratio validation, or agreement with independent free energy estimates. If so, the paper likely holds up. The abstract is well written and the authors clearly know the system. The concern is about evidence, not competence.\n\nVerdict: send to peer review. A serious referee should push for evidence of unbiasedness. If the full text checks that, this is a solid methods paper for computational chemists working on flexible complexes with multiple binding sites. If not, the speed-up claim is just a speed-up claim.\n\nShould I cite it? Not yet—I need the full validation. Reading group? Maybe, as a case study of how convergence metrics can mislead if the move set breaks detailed balance.\n\nRecommendation: accept for peer review, with the explicit ask to verify the swapping scheme's correctness.","headline":"Plausible extension of MT-REXEE with a useful case study, but the abstract can't support the unbiased-convergence claim; peer review should demand a detailed-balance check.","tokens_in":1466,"tokens_out":2153,"would_cite":false,"duration_ms":25119,"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":"A multiple-topology replica-exchange method uses ligand chain length as a route between alternate binding pockets, yielding faster convergence of relative binding free energies.","keywords":["binding free energy","alchemical free energy","replica exchange","expanded ensemble","enhanced sampling","FabB/ACP","multiple topology","alternate binding pockets"],"falsifier":"Run a long unbiased simulation of the FabB/ACP complex and measure the occupancy of the two binding pockets as a function of acyl chain length; if there is no chain length at which the ligand substantially occupies both pockets, the MT-REXEE swapping mechanism cannot work. Alternatively, apply MT-REXEE to a pair of binding pockets whose separation is insensitive to chain length and observe no convergence improvement.","tokens_in":719,"feed_emoji":"🧪","tokens_out":5480,"duration_ms":51977,"temperature":0.7,"pith_summary":"The paper argues that alchemical free-energy calculations often fail on flexible complexes because high barriers prevent ligands from visiting alternative binding poses within simulation time. It presents MT-REXEE, a multiple-topology replica-exchange expanded-ensemble approach in which the ligand is adaptively grown and shrunk along the alchemical chain-length dimension. In the FabB/ACP system, acyl chains bound in two distinct pockets can swap between pockets transiently at certain intermediate chain lengths; the method uses that transient overlap to let every replica visit both pockets, eliminating the need to define a collective variable that 'knows' the pockets in advance. The measured result is significantly faster convergence of relative binding free energies between the two pockets compared with standard approaches.","feed_headline":"Growing acyl chains lets simulations visit two binding pockets","feed_subtitle":"Chain length serves as the route between pockets, making binding free energy estimates converge faster.","key_machinery":"MT-REXEE (multiple topology replica exchange of expanded ensemble): a replica-exchange scheme in which each replica's Hamiltonian is an expanded ensemble over the ligand's alchemical coupling parameter, here the acyl chain length, and configurations are exchanged between replicas at different chain lengths. The key work it does is converting the physically slow transition between pockets into a fast reversible move along the alchemical dimension, so that growing or shrinking the chain visits the chain lengths where the ligand spontaneously switches pockets.","core_discovery":"The central claim is that MT-REXEE can overcome high free-energy barriers for flexible ligand–protein systems by using the alchemical progress parameter itself, the ligand's chain length, as a route between otherwise disconnected configurational states. In the FabB/ACP complex, the acyl chain grows while covalently attached to ACP, and at intermediate lengths the chain has nonzero probability of being found in either of two FabB pockets. By running replicas at different chain lengths and exchanging configurations among them, each simulation can access both pockets even when the physical barrier between them is high. This removes the requirement, inherited by collective-variable methods, of s","pith_inferences":["If ligand chain length is a sufficient reaction route between any pair of binding sites whose preference changes with ligand size, MT-REXEE could serve as a general proxy for an unknown collective variable, not just for this system.","A testable diagnostic follows: for a ligand whose two binding pockets are not connected by any physically accessible chain length, the method will not accelerate transitions; this could be checked by mutating the pocket or switching the ligand chemistry.","The alchemical dimension might also be coupled to protein conformational changes, potentially sampling coupled pocket-opening events, though the paper only demonstrates ligand-pose exchange."],"forward_implications":["For flexible complexes with high free-energy barriers, MT-REXEE can recover unbiased sampling of alternate binding modes without user-specified collective variables.","Relative binding free energies between kinetically separated pockets can be computed with faster convergence, reducing simulation time.","The open-source implementation makes the approach straightforward to apply to other ligand–protein systems with similar hidden barriers.","The method extends alchemical free-energy machinery to covalently attached ligands whose chain length can be grown or shrunk, a setting where conventional alchemical mutations are awkward."],"supporting_citations":[],"fun_headline_variants":["MT-REXEE uses chain length to bridge binding pockets","Growing acyl chains improves free energy sampling","Chain length as a route between two binding pockets","Enhanced sampling: chain growth enables pocket hopping"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"The enhanced sampling only works if, at some intermediate chain lengths, the ligand has a non-negligible probability of being found in both pockets, so that transitions between the pockets can be visited; if chain length does not open a route between the pockets, the speedup disappears.","fun_headline_variants_meta":{"raw":{"variants":["MT-REXEE uses chain length to bridge binding pockets","Growing acyl chains improves free energy sampling","Chain length as a route between two binding pockets","Enhanced sampling: chain growth enables pocket hopping"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000487,"raw_usage":{"total_tokens":2264,"prompt_tokens":796,"completion_tokens":1468,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":540,"completion_tokens_details":{"reasoning_tokens":1421}},"tokens_in":540,"tokens_out":1468,"duration_ms":11057,"temperature":1.0,"reasoning_tokens":1421,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T22:34:00.165593+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run a long unbiased simulation of the FabB/ACP complex and measure the occupancy of the two binding pockets as a function of acyl chain length; if there is no chain length at which the ligand substantially occupies both pockets, the MT-REXEE swapping mechanism cannot work. Alternatively, apply MT-REXEE to a pair of binding pockets whose separation is insensitive to chain length and observe no convergence improvement.","supporting_citations":[],"review_version":1}