- Rating
- 1
- Confidence
- 4
- Informativeness
- 1
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
COMPOSE combines an exact molecular graph-rewrite executor, a frozen task-agnostic reference law Rθ over canonical molecular successors, structured multi-edit programs, and task-specific controllers. It is evaluated on fragment-constrained generation, ZINC-QED editing, PMO-1K black-box optimization, and protein-specific lead optimization.
Critical strengths
- A coherent executable molecular-editing framework.
The exact executor constrains committed edits to a declared valid molecular graph space. Canonical-successor probabilities and multi-step programs form a reusable system
- Measurable benefits from learned transition preferences.
Held-out successor NLL decreases from 5.37 to 3.90, and learned edit preferences improve superstructure quality from 25.40% to 39.03%. Locked-panel selection studies offer additional evidence of utility
Critical weaknesses
1. The distinctive novelty and value of the learned process remain unestablished.
Autoregressive molecular graph generation with valence constraints (e.g., GraphAF), sequential editing and search (MIMOSA, MARS), reusable pretrained diffusion generators (GenMol), and goal guidance of frozen models (Value Matching) are established directions.
Canonical successors and structured programs may be a useful combination, but matched mechanism-level comparisons against strong editing and diffusion baselines are missing. Consequently, the paper does not establish what unique capability is enabled by this modeling paradigm.
Table 6 mainly evaluates prediction of constructed teacher transitions; Appendix F.2 states that ring-edit supervision comes entirely from synthetic walks. Free-running distribution and chemical-space diagnostics (e.g., FCD, novelty, scaffold coverage) are absent.
Moreover, although the paper emphasizes generation trajectories, coordinated edits, and retargeting, it provides almost no visualization of actual molecular transformations. This makes it difficult to verify whether the learned process exhibits distinctive behavior beyond conventional molecular search. The claimed 100% validity follows the declared structural guards; it does not establish synthetic feasibility or drug activity.
2. The optimization gains are insufficiently attributed, and benchmark comparability is limited.
The learned-law ablation is largely restricted to superstructure generation; PMO/T4 lack a controlled intervention on Rθ with the executor, programs, and controller fixed. T4 selects archive parents without score ranking, making the practical roles of Rθ and docking feedback unclear.
The structured-proposer ablation has only 14/18 completed pairs and simultaneously changes multiple proposal factors, preventing isolation of dependency scheduling (Appendix H.6).
On PMO, final Top-10 averages 0.563 versus GenMol's 0.542, but AUC-Top10 is 0.482 versus 0.537. Excluding the two isomer tasks leaves an approximately −0.015 average difference over the other 20 objectives.
The QED comparison uses different source panels and K=8 versus 20 candidates. T4 baselines were not rescored using the same docking pipeline, while the paper reports median repeat-docking variation of 0.70 kcal/mol. Eleven of the 23 nominal wins in Table 3 have margins below that scale.
These results do not yet substantiate broad superiority of the proposed modeling paradigm (Secs. 3.3–3.5; Appendices G, I).
Questions for authors
1. Novelty and comparisons.
What concrete capability does COMPOSE provide beyond GraphAF, MIMOSA/MARS, GenMol, and Value Matching? Can the key claim be tested with matched constraints, controllers, and budgets against strong graph-edit and diffusion baselines?
2. Causal attribution of the reference process and structured programs.
How do PMO and T4 change when Rθ is replaced by a uniform or empirical-family law while other components remain fixed? Where exactly do T4 proposals use Rθ and observed docking scores? Can the full 18/18 program ablation be completed with dependency scheduling isolated?
3. Benchmark comparability and optimization efficiency.
Can QED editing be evaluated on the same source panel with matched K, and can T4 baseline molecules be rescored consistently? Please provide PMO anytime Top-10 curves, sensitivity to the two isomer objectives, and proposal/compute cost per useful oracle evaluation.
4. Verifiable evidence of the actual generative process.
Can the authors show real molecule-by-molecule editing trajectories, highlighting changed atoms/bonds and transition probabilities, as well as dependency-aware multi-edit programs?
Matched visual examples comparing continued editing after a goal switch against restarting from the initial molecule would be particularly informative. Representative random successes and failures, together with free-running chemical-space diagnostics, would also help establish whether the claimed reusable process exhibits meaningful and distinctive behavior rather than merely producing competitive endpoint scores.
7 comments
Q2. Contributions of the reference process and structured programs
How do PMO and T4 change when Rθ is replaced by a uniform or empirical-family law while other components remain fixed? Where exactly do T4 proposals use Rθ and observed docking scores? Can the full 18/18 program ablation be completed with dependency scheduling isolated?
We appreciate the request to examine these mechanisms directly. We completed all 18 paired program comparisons and added an intervention on created-atom dependencies. To broaden the evaluation, we also compared learned, uniform, and empirical editing across all four fragment tasks. We report those results first, followed by the program and PMO experiments. We explain the T4 implementation in the next comment.
1. Learned transition preferences. For the uniform control, we sampled admissible canonical successors uniformly. For the empirical control, we selected edit families using fixed training frequencies. We then sampled uniformly within each selected family. Both controls retained the same executor and task constraints.
| Task quality (%) ↑ | COMPOSE | Uniform | Empirical |
|---|---|---|---|
| Motif extension | 42.63 ± 0.38 | 3.10 | 9.53 |
| Superstructure | 39.03 ± 3.09 | 1.97 | 33.77 |
| Linker design | 31.53 ± 1.40 | 22.75 | 23.70 |
| Scaffold decoration | 36.70 ± 1.80 | 3.07 | 11.17 |
With the learned reference, we obtained the highest quality on all four tasks. For motif extension, quality reached 42.63%, compared with 9.53% for empirical editing and 3.10% for uniform editing. We thus observed substantial gains from learned preferences within the legal editing space. We conducted this ablation separately from the expanded GenMol V2 fragment comparison.
2. Structured construction and created-atom dependencies. To examine structured construction, we retained learned primitive editing and removed the programs. We obtained the following fragment-generation results.
| Task quality (%) ↑ | COMPOSE | Learned primitive editing |
|---|---|---|
| Motif extension | 42.63 ± 0.38 | 7.87 ± 0.12 |
| Linker design | 31.53 ± 1.40 | 15.43 ± 0.91 |
We then completed the Appendix H.6 experiment across six objectives and three seeds, giving 18/18 paired comparisons.
| Program configuration | Final Top-10 ↑ | AUC-Top10 ↑ |
|---|---|---|
| Structured programs | 0.6298 | 0.5139 |
| Uniform editing chains | 0.5581 | 0.4682 |
| Created-atom binding intervention | 0.5799 | 0.4941 |
For the binding intervention, we retained the broader program recipe. We changed how later transformations attach to atoms created earlier. This reduced final Top-10 from 0.6298 to 0.5799, with paired Wilcoxon p = 0.0182. We therefore find direct support for preserving created-atom dependencies during molecular construction. This intervention tests atom binding within programs. The uniform-chain comparison evaluates the broader contribution of structured construction.
3. PMO ablations (see Appendix G.8 for PMO anytime Top-10 ablation curves). We additionally evaluated uniform and empirical legal-edit chains across the PMO objectives. For the learned primitive control, we retained neural preferences and removed structured programs. These comparisons examine the combined contribution of transition preferences and proposal construction. In the free-running experiment, we isolate the reference probabilities under identical execution rules.
| Objective | COMPOSE | Uniform | Empirical | Learned primitive |
|---|---|---|---|---|
| Albuterol similarity | 0.778 | 0.567 | 0.650 | 0.494 |
| Amlodipine MPO | 0.571 | 0.521 | 0.509 | 0.447 |
| Celecoxib rediscovery | 0.286 | 0.193 | 0.270 | 0.149 |
| Deco hop | 0.586 | 0.582 | 0.590 | 0.560 |
| DRD2 | 0.969 | 0.828 | 0.987 | 0.029 |
| Fexofenadine MPO | 0.752 | 0.657 | 0.686 | 0.585 |
| GSK3B | 0.502 | 0.424 | 0.447 | 0.276 |
| Isomers C7H8N2O2 | 1.000 | 0.855 | 0.882 | 0.040 |
| Isomers C9H10N2O2PF2Cl | 0.874 | 0.836 | 0.833 | 0.682 |
| JNK3 | 0.314 | 0.165 | 0.213 | 0.060 |
| Median1 | 0.278 | 0.227 | 0.272 | 0.182 |
| Median2 | 0.189 | 0.141 | 0.138 | 0.127 |
| Mestranol similarity | 0.516 | 0.407 | 0.428 | 0.270 |
| Osimertinib MPO | 0.843 | 0.788 | 0.779 | 0.726 |
| Perindopril MPO | 0.503 | 0.387 | 0.406 | 0.324 |
| QED | 0.938 | 0.925 | 0.918 | 0.836 |
| Ranolazine MPO | 0.783 | 0.574 | 0.665 | 0.209 |
| Scaffold hop | 0.459 | 0.447 | 0.464 | 0.426 |
| Sitagliptin MPO | 0.220 | 0.346 | 0.161 | 0.231 |
| Thiothixene rediscovery | 0.388 | 0.216 | 0.237 | 0.194 |
| Troglitazone rediscovery | 0.308 | 0.240 | 0.257 | 0.192 |
| Zaleplon MPO | 0.420 | 0.381 | 0.416 | 0.224 |
We also compared methods on shared objective sets. On the set shared with uniform editing, we obtained 0.570 with COMPOSE versus 0.490. On the set shared with learned primitive editing, we obtained 0.544 versus 0.405. Across these analyses, we consistently observed gains from learned, structured proposals. The completed paired experiment further supports the specific contribution of created-atom dependencies raised in the review.
Q2 continued. Reference guidance and docking in T4
In T4, we construct shallow edits, anchored replacements, and route-complete programs through the executor. For reference-guided selection, we score the executed transitions with the frozen reference. We average their log-probabilities to reweight program selection. We then dock feasible candidates and update the archive. We sample archive parents without docking-score ranking, with lower sampling frequency for exhausted parents. Docking scores determine the best feasible endpoints. The reference-law interventions above quantify effects in free-running walks and fragment generation.
Q3. Evaluation and efficiency
Can QED editing be evaluated on the same source panel with matched K, and can T4 baseline molecules be rescored consistently? Please provide PMO anytime Top-10 curves, sensitivity to the two isomer objectives, and proposal/compute cost per useful oracle evaluation.
1. Candidate allowance on a fixed QED panel. To measure the effect of candidate allowance, we evaluated the same 800 held-out sources using fixed COMPOSE candidate sequences. We scored the first eight or twenty candidates per source and counted missing positions as failures. Success requires QED ≥ 0.9 and source Tanimoto similarity ≥ 0.4.
| Candidate allowance | Successful sources | Success rate |
|---|---|---|
| K = 8 | 409 / 800 | 51.13% |
| K = 20 | 487 / 800 | 60.88% |
With twenty candidates, we improved success by 9.75 percentage points (95% source-bootstrap CI [7.75, 11.88]). This paired comparison quantifies the effect of candidate allowance within COMPOSE. For GrIDDD, the submission reports results from its published source panel.
2. T4 results and new baseline rescoring. We report all 30 settings across five targets, three starting leads, and two similarity floors. Feasible molecules satisfy QED ≥ 0.6, SA ≤ 4, and source Tanimoto similarity ≥ δ. C, I, and G denote COMPOSE, InVirtuoGen, and GenMol. Column labels specify δ. Docking scores are kcal/mol, with lower values preferred. NR denotes an unreported result.
| Target | Start | C .4 | I .4 | G .4 | C .6 | I .6 | G .6 |
|---|---|---|---|---|---|---|---|
| PARP1 | -7.3 | -11.9 | -14.1 | -10.6 | -13.6 | -12.3 | -10.4 |
| -7.8 | -12.3 | -13.4 | -11.0 | -12.7 | -11.7 | -9.7 | |
| -8.2 | -14.1 | -9.0 | -11.3 | -11.3 | -10.7 | -9.2 | |
| FA7 | -6.4 | -9.3 | -8.4 | -8.4 | -6.4 | -7.7 | -7.3 |
| -6.7 | -9.4 | -8.9 | -8.4 | -7.8 | -7.5 | -7.6 | |
| -8.5 | -9.6 | -8.0 | NR | -8.5 | -7.4 | NR | |
| 5HT1B | -4.5 | -12.4 | -13.3 | -12.9 | -13.3 | -12.4 | -12.1 |
| -7.6 | -11.9 | -12.0 | -12.3 | -12.2 | -12.0 | -12.0 | |
| -9.8 | -12.5 | -10.9 | -11.6 | -10.8 | -10.6 | -10.5 | |
| BRAF | -9.3 | -12.2 | -10.1 | -10.8 | -9.0 | -9.7 | NR |
| -9.4 | -10.5 | -10.8 | -10.8 | -11.0 | -10.4 | -9.7 | |
| -9.8 | -10.9 | -10.6 | -10.6 | -11.5 | -10.3 | -10.5 | |
| JAK2 | -7.7 | -10.4 | -10.2 | -10.2 | -10.6 | -9.7 | -9.3 |
| -8.0 | -11.3 | -10.5 | -10.0 | -10.8 | -10.4 | -9.4 | |
| -8.6 | -10.7 | -10.2 | -9.8 | -10.8 | -10.3 | NR |
With COMPOSE, we obtained the best reported docking score in 23/30 settings using 250 evaluations per lead, compared with 1,000 for the baselines. We retain the published benchmark scores in this table.
To examine scoring variability directly, we rescored 90 released InVirtuoGen endpoints three times each using the COMPOSE pipeline. This produced 270 new docking measurements.
| Within-molecule docking spread | kcal/mol |
|---|---|
| Median | 0.10 |
| 90th percentile | 1.30 |
| Maximum | 6.50 |
For these endpoints, we measured a median spread of 0.10 kcal/mol. We report the 90th percentile and maximum to describe the larger variations across individual molecules. The earlier repeat-docking analysis in the submission gave a median of 0.70. Here, we characterize the released baseline endpoints under our pipeline. The lead-by-lead table above retains the published scores.
3. Useful progress per oracle call. To assess evaluation efficiency, we compared seven methods across 16 matched PMO objectives. Each method received 1,000 calls. We counted an evaluation as useful when it improved the running Top-10 after the 16 initialization molecules.
| Method | Calls/improvement ↓ | Relative calls | COMPOSE wins | p-value |
|---|---|---|---|---|
| COMPOSE | 9.03 | 1.00x | — | — |
| GenMol V2 | 10.47 | 1.16x | 13/16 | 1.0e-2 |
| GenMol | 10.70 | 1.18x | 14/16 | 1.4e-3 |
| Graph-GA | 15.87 | 1.76x | 15/16 | 1.5e-7 |
| MARS | 21.87 | 2.42x | 16/16 | 1.5e-8 |
| MIMOSA | 22.36 | 2.48x | 16/16 | 3.0e-8 |
| GraphAF | 24.60 | 2.73x | 16/16 | 1.5e-8 |
With COMPOSE, we required a median of 9.03 calls per useful improvement. GenMol required 1.16–1.18 times as many calls, while graph-based comparators required 1.76–2.73 times as many. COMPOSE won on 13–16 of 16 objectives across baselines. This demonstrates more efficient use of objective evaluations. We measure oracle-call cost here, with proposal-generation time and total runtime as separate quantities.
We thank the reviewer for the careful assessment and specific suggestions. To address the concerns, we conducted new experiments on the learned reference, dependent edits, and continued optimization after a goal switch. We also expanded the baseline comparison to eight methods and completed all 18 paired program experiments. We explain the contribution first, then address each experimental question in turn.
Q1. What capability does COMPOSE add?
With COMPOSE, we learn probabilities over executable changes between complete molecular graphs. We can compose these changes into programs with explicit dependencies between edits. For example, we can insert an atom and later close a ring through that same atom. During execution, we track atom identities across steps and verify every intermediate molecule. We aggregate equivalent edits into one probability for each distinct molecular outcome.
The reviewer correctly identifies established work on autoregressive graph construction [1], molecular sampling [2,3], and guidance of frozen generators [4,5]. Our contribution combines canonical transition probabilities with executable programs that preserve dependencies between molecular edits. This makes it possible to continue editing an existing molecule under a new objective using the same frozen reference. We test that capability directly in the goal-switch experiments described in our final comment.
To compare optimization performance, we evaluated eight methods on 22 PMO objectives with 1,000 charged calls per run. With COMPOSE, we obtained the highest mean final Top-10 of 0.567, compared with 0.563 for GenMol and 0.555 for GenMol V2. We retained each baseline's optimization procedure under the common budget. For the mechanism comparisons, we separately tested the reference probabilities and program construction within COMPOSE.
W1. Free-running behavior beyond teacher-transition prediction
We agree that examining free-running molecules is important for evaluating the learned process. To address this, we generated four eight-edit trajectories from each of 512 held-out molecules, producing 2,048 endpoints per arm. We used identical starting molecules, execution rules, and stopping criteria for learned and uniform editing. We disabled optimization controllers and oracle feedback in both arms. We then independently replayed all 32,768 executed edits.
For FCD, we compared both endpoint sets with the same separate panel of 2,048 held-out molecules. We also measured recovery of the panel's 458 distinct scaffolds. Using the learned probabilities reduced FCD from 14.35 to 2.77 and increased scaffold coverage from 15.28% to 25.11%. We also obtained higher QED and lower SA scores under the same validity rules. These improvements demonstrate the value of learned transition preferences during free-running molecular editing.
We found zero endpoint overlap with the available training-source bank in both arms. The starting molecules were also held out. These are source-conditioned walks, with synthetic ring-edit supervision as described in Appendix F.2. We therefore evaluate the resulting molecular distributions beyond teacher-transition prediction. QED, SA, and docking are computational property measurements.
To further examine reuse, we compared continuation after a goal switch with restarting from the original molecule. Continuation performed better in 35/40 potency-first and 29/40 developability-first comparisons under the same controller and three-edit allowance. We provide the molecular examples and paired confidence intervals in the final comment.
[1] Shi et al., GraphAF, ICLR 2020. [2] Fu et al., MIMOSA, AAAI 2021. [3] Xie et al., MARS, ICLR 2021. [4] GenMol and V2. [5] Jensen et al., Value Matching, ICLR 2026.