Pith. sign in

REVIEW 4 major objections 5 minor 24 references

Uncertainty-Aware Metabolic Stability Prediction with Dual-View Contrastive Learning

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read TrustworthyMS, a graph neural network that pairs atom- and bond-level views through contrastive learning, claims to outperform state-of-the-art metabolic stability predictors on both classification and regression while supplying…

desk verdict Solid incremental paper with a plausible classification result; the regression branch has a target-scaling mismatch that makes the headline HL gains unverified until the code is checked. read the letter →

arxiv 2506.00936 v1 pith:TSOI4L2I submitted 2025-06-01 cs.LG cs.AIq-bio.QM

Peijin Guo , Minghui Li , Hewen Pan , Bowen Chen , Yang Wu , Zikang Guo , Leo Yu Zhang , Shengshan Hu
show 1 more author
Shengqing Hu
This is my paper · ORCID
classification cs.LGcs.AIq-bio.QM
keywords metabolicstabilitypredictioncontrastivelearninggraphneuralnetworkuncertaintyquantificationBeta-Binomialevidencedrugdiscoverybond-leveltopologyout-of-distributiongeneralization
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper aims to establish a metabolic stability predictor that gets two things at once: higher accuracy than existing graph-based models and a confidence score for each prediction. The proposed architecture, TrustworthyMS, combines a molecular graph whose nodes include atom-bond-atom triplets, a contrastive loss that aligns the atom view with the bond view, and a Beta-Binomial evidential head that outputs both a prediction and an uncertainty estimate. On the HLM classification benchmark it reports a Matthews correlation coefficient of 0.622, and on the half-life regression benchmark a root mean squared error of 0.091, both better than the previous state of the art. If the claims hold, drug-discovery teams could rank candidate molecules by predicted stability while knowing which predictions to trust.

What carries the argument

The load-bearing object is the dual-view graph pair. Molecular graph topology remapping first creates bond-centric nodes $v^r_{ij} = f_{\mathrm{node}}(v_i \oplus e_{ij} \oplus v_j)$ from an atom, its bond, and the neighboring atom, then links those nodes into a second graph when two bonds share an atom. A GIN encoder with centering-and-scaling anti-smoothing normalization produces embeddings for both views, and the contrastive loss $L_{\mathrm{CL}}$ of Eq. 8 aligns the molecular-topology view with the bond-interaction view. The evidential head uses Beta-Binomial conjugacy: classification evidence $(e^+, e^-)$ maps to Beta parameters $\alpha = e^+ + 1$, $\beta = e^- + 1$, giving belief, disbelief, and uncertainty masses by Eq. 11, while regression generates $(\alpha, \beta)$ through sigmoid networks and predicts the Beta mean while penalizing variance in Eq. 17. The joint objective $L = L_G + L_{G^r} + \lambda L_{\mathrm{CL}}$ is trained end to end.

What would settle it

Inspect the actual preprocessing values: if any z-score-normalized half-life target is at or below 0 or at or above 1, then Eq. 17 is not a proper Beta-expectation loss for those targets. A direct check is to rerun the regression experiments with targets min-max scaled into $(0,1)$ and compare RMSE and Spearman correlation; if the reported gains vanish, the evidential loss was doing the work, and if they persist, the representation and contrastive components carry the improvement.

Watch

Extended reading notes

Core claim

The central claim is that representing a molecule as two aligned graphs—the original atom graph $G$ and a bond-centric remapped graph $G^r$ built from atom-bond-atom triplets—lets the model capture interactions that atom-only message passing misses, and that attaching a Beta-Binomial evidential head turns each output into a Beta distribution whose mean is the prediction and whose spread is the uncertainty. The paper reports that this architecture beats existing methods on both tasks: on HLM classification the MCC is 0.622 versus 0.601 for MS-BACL, and on half-life regression the RMSE is 0.091 versus 0.111 with Spearman correlation 0.833 versus 0.652. It also reports an out-of-distribution MCC of 0.615 versus 0.588 when the HLM-trained model is tested on structurally dissimilar compounds. The same uncertainty estimates are used as a filter: as the acceptance threshold tightens, reported accuracy on HLM rises from about 83 percent to 94 percent, and the regression correlation rises from about 0.83 to 0.90.

Load-bearing premise

The regression loss in Eq. 17 computes the expectation of squared error under a Beta distribution and therefore assumes the normalized half-life targets lie strictly between 0 and 1, but Section 4.1 says the half-life dataset was standardized via z-score normalization, which can produce negative values and values above 1; if the dataset was not rescaled into $(0,1)$, that loss is not a valid likelihood for those targets.

Editorial extensions

If this is right

  • Bond-level topology remapping is what lifts classification: removing it drops MCC on HLM from 0.622 to 0.613.
  • Contrastive alignment is what most helps regression: removing dual-view contrastive learning raises RMSE from 0.091 to 0.097 and lowers the Spearman correlation from 0.833 to 0.818.
  • Uncertainty thresholds give a deployable reliability filter: accepting only the highest-confidence predictions raises HLM accuracy to about 94 percent and half-life correlation to about 0.90.
  • The model generalizes to out-of-distribution scaffolds better than the previous state of the art: OOD MCC is 0.615 versus 0.588 for MS-BACL.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the z-score normalization was not rescaled into $(0,1)$, the evidential regression loss is misspecified, and the reported regression advantage could be carried by the contrastive and remapping components; rerunning with min-max scaling would separate these effects.
  • The same Beta-Binomial head could be attached to other ADME endpoints such as clearance or permeability, where calibrated confidence matters as much as point accuracy.
  • The dual-view contrastive objective is not tied to atom-bond duality; aligning the same encoder across 3D conformer views or fingerprint views would test whether the improvement is specific to bond remapping or general to multi-view regularization.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes TrustworthyMS, a graph neural network framework for metabolic stability prediction with three claimed contributions: molecular graph topology remapping that augments atom-centric graphs with bond-centric nodes, dual-view contrastive learning between the molecular and bond views, and Beta-Binomial evidential uncertainty quantification. The authors evaluate on HLM classification (5,876 compounds), Half-Life regression (656 compounds), and a composite OOD set, reporting improved MCC/F1 over MS-BACL in classification, improved RMSE/P in regression, and an OOD robustness improvement. The paper also includes ablations, a parameter sensitivity study, an uncertainty-threshold reliability analysis, and a Shapley-value case study.

Significance. If the results hold, the main contribution is a practical integration of bond-aware graph augmentation, contrastive learning, and evidential uncertainty estimation for ADME-relevant property prediction. The release of code and the use of established benchmarks (HLM, Half-Life, OOD sets) are strengths, and the uncertainty-threshold analysis gives a useful direction for reliability-based screening. However, the central 'outperforms current state-of-the-art' claim is currently weakened by an apparent mismatch between the stated z-score normalization of the regression targets and the Beta-mean predictor, which is confined to (0,1), and by several undefined or internally inconsistent numerical claims. The uncertainty quantification contribution is not benchmarked against standard UQ baselines, so the calibration claim is not yet quantitatively supported.

major comments (4)
  1. [§4.1 and §3.4 (Eqs. 13, 15, 17)] Section 4.1 states that the Half-Life dataset is 'standardized via z-score normalization.' In Eq. 13, α and β are produced by sigmoid outputs plus ε, so both lie in (ε, 1+ε), and the point prediction in Eq. 15, α/(α+β), is confined to the open interval (0,1). A z-scored target necessarily takes negative values (the mean is 0 after centering), so the described model cannot represent part of the target distribution. Equation 17 explicitly assumes 'normalized targets y_i ∈ (0,1),' which is inconsistent with z-score normalization. This is load-bearing because the regression margin over MS-BACL (RMSE 0.091 vs 0.111) is one of the strongest claimed improvements. The authors must clarify whether the actual preprocessing is min-max scaling (a text error) or z-score normalization (a model misspecification), and adjust the text, equations, or results accordingly.
  2. [Abstract, §4.2, Conclusion] The claim of a '46.1% improvement in robustness on out-of-distribution (OOD) data' is undefined and does not match any clean comparison in Table 3. The direct comparison to the previous SOTA method MS-BACL is MCC 0.615 vs 0.588, a 4.6% relative improvement, while comparisons to GAT (0.414) or MGCN (0.447) give 48.6% or 37.6%, respectively. No definition of 'robustness' is provided, yet the same 46.1% figure is repeated in the Conclusion. Please define the metric, show the calculation, or remove the claim.
  3. [§4.2 (Tables 1–3)] The paper asserts that TrustworthyMS 'outperforms current state-of-the-art methods,' but no significance tests are reported and several key differences are within one standard deviation of the baselines (e.g., HLM MCC 0.622±0.034 vs 0.601±0.053; F1 0.866±0.012 vs 0.863±0.018). Moreover, on the OOD set the AUC is lower than MS-BACL (0.862 vs 0.897), and on HLM classification the AUC is tied (0.873). The text acknowledges the lower OOD AUC but still frames the overall result as SOTA. Please report paired significance tests (e.g., Wilcoxon signed-rank or bootstrap) or temper the superiority claim to the specific metrics that are actually improved.
  4. [§4.2, 'Uncertainty-Aware Prediction Reliability'] The uncertainty quantification is evaluated only through thresholding on the model's own uncertainty scores, with no comparison to standard uncertainty quantification baselines such as Monte Carlo dropout, deep ensembles, or other evidential methods, and no calibration curve or calibration metric (e.g., reliability diagram or expected calibration error). Since uncertainty quantification is a central contribution of the paper, the claim of 'calibrated uncertainty estimation' needs a quantitative benchmark against existing UQ approaches, not only an internal consistency check.
minor comments (5)
  1. [§4.2] Several percentage changes are computed inconsistently: the '9.8% MCC gain' from 0.564 to 0.622 is actually 10.3%, the '9.0% improvement versus CMMS-GCL' from 0.566 to 0.622 is 9.9%, and the '48.2% improvement over basic GNNs (D-MPNN: 0.218)' should be approximately 182% under the formula used elsewhere. Please recalculate and standardize the reporting of relative improvements.
  2. [Abstract and §4.2] The abstract and conclusion refer to a dataset of '10,031 compounds,' but the described datasets sum to 7,142 compounds (HLM 5,876 + HL 656 + OOD 610). Please reconcile this discrepancy.
  3. [§3.2 and Eq. (8)] The contrastive loss in Eq. (8) is written on graph encoder outputs z, while Section 3.2 introduces projection heads g(·) producing p_G and states that the contrastive loss operates on the encoder outputs. The architecture figure also shows projection heads feeding the contrastive loss. Please clarify whether the loss uses z or p.
  4. [§4.1, §4.2, Fig. 1] There are several typographical issues: 'with to ensure distributional fairness' in §4.1, 'regresion' in §4.2, 'jontly' in the Fig. 1 caption, and '0.825±0.15' in Table 4 should presumably be '0.825±0.015.'
  5. [§3.2 and Fig. 3] The anti-smoothing scaling factor s is fixed at 1×10⁻⁶, but the sensitivity analysis in Fig. 3 sweeps only the contrastive weight λ. Since s affects the feature magnitudes substantially, a brief discussion or sensitivity check for s would help.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported gains come from held-out benchmark comparisons, and the architecture/loss derivations are self-contained.

full rationale

The paper's central claim is empirical: TrustworthyMS outperforms baselines on in-distribution, OOD, and ablation benchmarks. These results are obtained by training on training folds and evaluating on test folds, so they are not fitted inputs renamed as predictions. The Beta-Binomial evidential loss in Eq. 16-17 is a training objective derived from a modeling choice, not a quantity that presupposes the reported performance. The uncertainty-threshold analysis (Fig. 4) uses the model's own uncertainty scores to show that lower predicted uncertainty correlates with higher accuracy; this is a self-consistency check rather than a circular derivation of the SOTA claim. The manuscript contains self-citations (refs. [5], [10], [11]) but they are prior related work by the same group and are not load-bearing for the core result. The only substantive concern is an internal consistency issue: Section 4.1 states that the Half-Life targets are z-score normalized, while Eq. 17 and Eq. 15 assume targets in (0,1) and predictions confined to (0,1). That mismatch, if real, is a correctness or reproducibility problem, not a circularity problem, because it does not make the reported prediction equivalent to the model's inputs by construction. Hence the circularity score is 0.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

No new physical entities are introduced. The ledger is dominated by architectural hyperparameters and domain assumptions about why bond-centric graphs should help. The most fragile entry is the regression target support assumption, which conflicts with the stated preprocessing.

free parameters (6)
  • contrastive loss weight lambda = not stated; swept over [0.1, 0.9]
    Eq. 18 combines L_G + L_Gr + lambda L_CL; Section 4.2 sweeps lambda but does not report the value used for final results.
  • anti-smoothing scaling factor s = 1e-6
    Fixed constant in Eq. 6, chosen by hand.
  • contrastive temperature tau = 0.5
    Temperature in Eq. 8, chosen by hand.
  • evidence epsilon = 1e-6
    Numerical stabilizer in Eq. 13, chosen by hand.
  • Dirichlet prior constant K = 2
    Prior constant in Eq. 11 for binary classification, standard but chosen.
  • OOD Tanimoto dissimilarity threshold = <0.35
    Data split criterion following prior work [20], affects OOD evaluation.
assumptions (5)
  • domain assumption Beta-Binomial conjugacy with K=2 maps network evidence to belief, disbelief, and uncertainty masses (Eq. 11).
    Section 3.3 assumes this subjective logic mapping is valid for neural evidence outputs; the prior K is fixed rather than learned or justified.
  • domain assumption The bond-centric remapped graph Gr and its relation edges capture pharmacophore and steric information relevant to metabolic stability.
    Section 3.1 states this motivation; it is asserted, not derived, and the improvement over MS-BACL is used as indirect evidence.
  • domain assumption Contrastive alignment between the two views maximizes mutual information and improves representation robustness.
    Section 3.2 interprets L_CL via mutual information; no theoretical guarantee is given for this particular dual-graph setting.
  • standard math Standard GIN message passing and hybrid pooling produce sufficient graph representations.
    Eq. 4 relies on usual GNN assumptions.
  • ad hoc to paper For regression, targets are assumed to lie in (0,1) for the Beta expectation loss.
    Eq. 17 requires y_i in (0,1), but Section 4.1 says z-score normalization; this is an inconsistency in the paper's own setup.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Uncertainty-Aware Metabolic Stability Prediction with Dual-View Contrastive Learning." pith.science (2026). https://pith.science/paper/TSOI4L2I

@misc{pith2026250600936,
  author       = {Pith},
  title        = {Pith review of: Uncertainty-Aware Metabolic Stability Prediction with Dual-View Contrastive Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TSOI4L2I}},
  note         = {Machine review of arXiv:2506.00936}
}
read the original abstract

Accurate prediction of molecular metabolic stability (MS) is critical for drug research and development but remains challenging due to the complex interplay of molecular interactions. Despite recent advances in graph neural networks (GNNs) for MS prediction, current approaches face two critical limitations: (1) incomplete molecular modeling due to atom-centric message-passing mechanisms that disregard bond-level topological features, and (2) prediction frameworks that lack reliable uncertainty quantification. To address these challenges, we propose TrustworthyMS, a novel contrastive learning framework designed for uncertainty-aware metabolic stability prediction. First, a molecular graph topology remapping mechanism synchronizes atom-bond interactions through edge-induced feature propagation, capturing both localized electronic effects and global conformational constraints. Second, contrastive topology-bond alignment enforces consistency between molecular topology views and bond patterns via feature alignment, enhancing representation robustness. Third, uncertainty modeling through Beta-Binomial uncertainty quantification enables simultaneous prediction and confidence calibration under epistemic uncertainty. Through extensive experiments, our results demonstrate that TrustworthyMS outperforms current state-of-the-art methods in terms of predictive performance.

Figures

Figures reproduced from arXiv: 2506.00936 by the authors.

Figure 1
Figure 1. TrustworthyMS framework architecture. The model integrates three compo￾nents: (a) molecular graph topology remapping constructs dual atom- and bond-centric representations, (b) dual-view contrastive learning jontly optimizes molecular topol￾ogy and bond-interaction embeddings via feature alignment, (c) evidential uncertainty quantification predicts metabolic stability with calibrated confidence. 3.1 Molecular Graph … view at source ↗
Figure 2
Figure 2. Graph encoder architecture integrating GIN layers, anti-smoothing normaliza￾tion, and hybrid pooling 3.2 Dual-View Contrastive Learning Our framework implements dual-view contrastive learning to align molecular topology with bond-interaction semantics through graph encoding and projec￾tion. As shown in [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Parameter sensitivity analysis of contrastive learning coefficient λ. The model exhibits robustness both o the HLM classification (left) and HL regression (right). Shaded regions denote 95% confidence intervals over 10 trials. Parameter Sensitivity Analysis. The contrastive learning coefficient λ in Eq. 18 governs the trade-off between discriminative evidence learning (via LCL) and predictive accuracy optimization. … view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Uncertainty-guided performance improvement. (a) HLM classification accuracy vs. retained sample ratio under decreasing uncertainty thresholds. (b) HL regression P vs. retention rate. Shaded regions denote 95% confidence intervals over 10 trials. Uncertainty-Aware Predi…
Figure 5
Figure 5. Figure 5: Multifaceted analysis of metabolic stability determinants. (A) Negative-impact functional group frequencies. (B) Smallest molecule in HLM, where S represents the length of the SMILES sequence. (C) Largest molecule in HLM. (D) Shortest half-life molecule in HL, with log…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

24 extracted references · 23 canonical work pages

  1. [1]

    Chemometrics and Intelligent Laboratory Systems257, 105292 (2025) 16 Authors Suppressed Due to Excessive Length

    Deng, S., Wu, Y., Ye, Z., Ouyang, D.: In silico prediction of metabolic stability for ester-containing molecules: Machine learning and quantum mechanical methods. Chemometrics and Intelligent Laboratory Systems257, 105292 (2025) 16 Authors Suppressed Due to Excessive Length

    work page 2025

  2. [2]

    Bioinformatics39(8), btad503 (2023)

    Du, B.X., Long, Y., Li, X., Wu, M., Shi, J.Y.: Cmms-gcl: cross-modality metabolic stability prediction with graph contrastive learning. Bioinformatics39(8), btad503 (2023)

    work page 2023

  3. [3]

    Drug metabolism reviews53(3), 459–477 (2021)

    Gajula, S.N.R., Nadimpalli, N., Sonti, R.: Drug metabolic stability in early drug discovery to develop potential lead compounds. Drug metabolism reviews53(3), 459–477 (2021)

    work page 2021

  4. [4]

    ChemMedChem 16(24), 3772–3786 (2021)

    Grebner, C., Matter, H., Kofink, D., Wenzel, J., Schmidt, F., Hessler, G.: Applica- tion of deep neural network models in drug discovery programs. ChemMedChem 16(24), 3772–3786 (2021)

    work page 2021

  5. [5]

    Multi-Modality Representation Learning for Antibody-Antigen Interactions Prediction

    Guo, P., Li, M., Pan, H., Huang, R., Xue, L., Hu, S., Guo, Z., Wan, W., Hu, S.: Multi-modality representation learning for antibody-antigen interactions predic- tion. arXiv preprint arXiv:2503.17666 (2025)

    work page Pith review arXiv 2025

  6. [6]

    Scientific Reports15(1), 7334 (2025)

    Isa, A.S., Uzairu, A., Umar, U.M., Ibrahim, M.T., Umar, A.B., Tabti, K., Mo- hammed, A.M.: In silico exploration of novel egfr-targeting compounds: integrative molecular modeling, docking, pharmacokinetics, and md simulations for advancing anti-cervical cancer therapeutics. Scientific Reports15(1), 7334 (2025)

    work page 2025

  7. [7]

    Springer Publishing Company, Incorporated (2018)

    Jsang, A.: Subjective Logic: A formalism for reasoning under uncertainty. Springer Publishing Company, Incorporated (2018)

    2018

  8. [8]

    Nature reviews Drug discovery3(11), 935–949 (2004)

    Kitchen, D.B., Decornez, H., Furr, J.R., Bajorath, J.: Docking and scoring in vir- tual screening for drug discovery: methods and applications. Nature reviews Drug discovery3(11), 935–949 (2004)

    work page 2004

Show all 24 references
  1. [9]

    Chemical Research in Toxicology35(9), 1614–1624 (2022)

    Li, L., Lu, Z., Liu, G., Tang, Y., Li, W.: In silico prediction of human and rat liver microsomal stability via machine learning methods. Chemical Research in Toxicology35(9), 1614–1624 (2022)

    2022

  2. [10]

    In: 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)

    Li, M., Guo, Z., Wu, Y., Guo, P., Shi, Y., Hu, S., Wan, W., Hu, S.: Vidta: Enhanced drug-target affinity prediction via virtual graph nodes and attention-based feature fusion. In: 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM). pp. 42–47. IEEE (2024)

    2024

  3. [11]

    Bioinformatics p

    Li, M., Shi, Y., Hu, S., Hu, S., Guo, P., Wan, W., Zhang, L.Y., Pan, S., Li, J., Sun, L., et al.: Mvsf-ab: Accurate antibody-antigen binding affinity prediction via multi-view sequence feature learning. Bioinformatics p. btae579 (2024)

    2024

  4. [12]

    iscience25(10) (2022)

    Mastropietro, A., Pasculli, G., Feldmann, C., Rodríguez-Pérez, R., Bajorath, J.: Edgeshaper: Bond-centric shapley value-based explanation method for graph neu- ral networks. iscience25(10) (2022)

    2022

  5. [13]

    Nucleic acids research47(D1), D930–D940 (2019)

    Mendez, D., Gaulton, A., Bento, A.P., Chambers, J., De Veij, M., Félix, E., Maga- riños, M.P., Mosquera, J.F., Mutowo, P., Nowotka, M., et al.: Chembl: towards di- rect deposition of bioassay data. Nucleic acids research47(D1), D930–D940 (2019)

    2019

  6. [14]

    Phar- maceutical research33, 433–449 (2016)

    Perryman, A.L., Stratton, T.P., Ekins, S., Freundlich, J.S.: Predicting mouse liver microsomal stability with“pruned” machine learning models andpublicdata. Phar- maceutical research33, 433–449 (2016)

    2016

  7. [15]

    International journal of molecular sciences19(4), 1040 (2018)

    Podlewska, S., Kafel, R.: Metstabon—online platform for metabolic stability pre- dictions. International journal of molecular sciences19(4), 1040 (2018)

    2018

  8. [16]

    Nature Communications16(1), 2156 (2025)

    Puumala, E., Nandakumar, M., Yiu, B., Stogios, P.J., Strickland, B.G., Zarnowski, R., Wang, X., Williams, N.S., Savchenko, A., Andes, D.R., et al.: Structure-guided optimization of small molecules targeting yck2 as a strategy to combat candida albicans. Nature Communications16...

    2025

  9. [17]

    Wiley interdisciplinary reviews: computational molecular science11(1), e1479 (2021) Title Suppressed Due to Excessive Length 17

    Renn, A., Su, B.H., Liu, H., Sun, J., Tseng, Y.J.: Advances in the prediction of mouse liver microsomal studies: from machine learning to deep learning. Wiley interdisciplinary reviews: computational molecular science11(1), e1479 (2021) Title Suppressed Due to Excessive Length 17

    2021

  10. [18]

    Bioinformatics38(2), 364–368 (2022)

    Ryu, J.Y., Lee, J.H., Lee, B.H., Song, J.S., Ahn, S., Oh, K.S.: Predms: a random forest model for predicting metabolic stability of drug candidates in human liver microsomes. Bioinformatics38(2), 364–368 (2022)

    2022

  11. [19]

    Journal of cheminformatics12, 1–14 (2020)

    Shah, P., Siramshetty, V.B., Zakharov, A.V., Southall, N.T., Xu, X., Nguyen, D.T.: Predicting liver cytosol stability of small molecules. Journal of cheminformatics12, 1–14 (2020)

    2020

  12. [20]

    Briefings in Bioinformatics25(3), bbae127 (2024)

    Wang, T., Li, Z., Zhuo, L., Chen, Y., Fu, X., Zou, Q.: Ms-bacl: enhancing metabolic stability prediction through bond graph augmentation and contrastive learning. Briefings in Bioinformatics25(3), bbae127 (2024)

    2024

  13. [21]

    Drug Discovery Today: Technologies37, 1–12 (2020)

    Wieder, O., Kohlbacher, S., Kuenemann, M., Garon, A., Ducrot, P., Seidel, T., Langer, T.: A compact review of molecular property prediction with graph neural networks. Drug Discovery Today: Technologies37, 1–12 (2020)

    2020

  14. [22]

    Signal Transduction and Targeted Therapy10(1), 1–56 (2025)

    Xiao, W., Jiang, W., Chen, Z., Huang, Y., Mao, J., Zheng, W., Hu, Y., Shi, J.: Advance in peptide-based drug development: delivery platforms, therapeutics and vaccines. Signal Transduction and Targeted Therapy10(1), 1–56 (2025)

    2025

  15. [23]

    Journal of medicinal chemistry 63(16), 8749–8760 (2019)

    Xiong, Z., Wang, D., Liu, X., Zhong, F., Wan, X., Li, X., Li, Z., Luo, X., Chen, K., Jiang, H., et al.: Pushing the boundaries of molecular representation for drug discovery with the graph attention mechanism. Journal of medicinal chemistry 63(16), 8749–8760 (2019)

    2019

  16. [24]

    In: Joint European Conference on Machine Learning and Knowledge Discovery in Databases

    Ye, Y., Zhou, J., Li, S., Xiao, C., Ying, H., Xiong, H.: Hierarchical structure- aware graph prompting for drug-drug interaction prediction. In: Joint European Conference on Machine Learning and Knowledge Discovery in Databases. pp. 36–

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.