Pith. sign in

REVIEW 4 major objections 5 minor 32 references

Data Leakage and Redundancy in the LIT-PCBA Benchmark

T0 review · 4 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A memorization baseline with no learnable parameters matches the median EF1% of state-of-the-art 3D encoders on LIT-PCBA, so the benchmark's scores measure scaffold memorization rather than generalization to novel chemotypes.

desk verdict The leakage audit is real, reproducible, and long overdue; the headline parity claim is not supported as stated. read the letter →

arxiv 2507.21404 v2 pith:23HOTS6V submitted 2025-07-29 cs.LG q-bio.QM

Amber Huang , Ian Scott Knight , Slava Naprienko This is my paper · ORCID
classification cs.LGq-bio.QM
keywords dataleakagevirtualscreeningbenchmarkLIT-PCBAmemorizationbaselinemolecularredundancyscaffoldenrichmentfactoraudit
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 tries to establish that LIT-PCBA, a widely used virtual-screening benchmark, is so riddled with duplicated and near-duplicated molecules across its training, validation, and query splits that published state-of-the-art scores do not measure generalization to new chemotypes. Its decisive experiment is a trivial memorization baseline: for each validation molecule, take the maximum ECFP4 Tanimoto similarity to any training active and rank by that score. With no learnable parameters, this baseline reaches a median raw enrichment factor of 4.15 at the 1% threshold, matching the median EF1% reported for the CHEESE 3D encoders under the same multi-ligand protocol. The audit also counts 2,491 2D-identical inactives shared between training and validation sets and more than 350 active analog pairs at Tanimoto similarity at least 0.6, arguing that the leakage is structural, not incidental. If the claim holds, published LIT-PCBA results should be reinterpreted as measuring scaffold memorization rather than screening skill.

What carries the argument

The load-bearing object is the memorization baseline: a 4096-bit ECFP4 fingerprint is computed for each molecule, stereochemistry is stripped so that stereo-agnostic duplicates count as identical, and each validation molecule is scored by its maximum Tanimoto similarity to any fingerprint in the set of training actives. The baseline has no learnable parameters and uses no physical modeling, so any enrichment it achieves must come from information the benchmark itself puts into both splits. The named comparison object is EF1%, the enrichment factor at the top 1% of the screened library, along with its normalized variant nEF1%, which divides raw EF1% by that target's theoretical maximum. The baseline's parity with the 3D encoders is the step that turns the leakage inventory into a claim about published results.

What would settle it

Run the released memorization baseline and the two CHEESE 3D encoders on the official LIT-PCBA splits under the multi-ligand max-pooling protocol and compare median raw EF1% values computed from code; if the CHEESE medians do not reproduce the transcribed values, the parity claim fails.

Watch

Extended reading notes

Core claim

The central discovery is that LIT-PCBA's validation set is not held out in any meaningful sense. Across 15 targets the paper finds stereo-agnostic duplicate molecules inside the same split and across splits, extensive analog overlap at ECFP4 Tanimoto similarity at least 0.6, and query sets dominated by near-identical ligands; in MTORC1, for example, nine of the eleven query entries reduce to three highly similar scaffolds. Because of this, a stereo-agnostic memorization baseline that simply max-pools ECFP4 similarity to training actives scores a median raw EF1% of 4.15, equal to the median reported for the CHEESE 3D encoders in the multi-ligand comparison, while selective reporting of the median rather than the mean, and failure to report normalized nEF1%, can invert which method appears superior. The paper concludes that nearly all published LIT-PCBA results, including zero-shot evaluations, are inflated by these artifacts, and that the benchmark cannot be repaired by de-duplication because analog leakage and low query diversity are systemic.

Load-bearing premise

The head-to-head parity claim depends on the assumption that the published CHEESE EF1% values, which the authors transcribed from figures, were produced under the same multi-ligand max-pooling protocol used for the memorization baseline, and those values are not supplied in a machine-readable form in this paper.

Editorial extensions

If this is right

  • Published LIT-PCBA enrichment factors and AUROC scores should be read as measuring scaffold memorization, not recovery of novel chemotypes.
  • Zero-shot evaluations are also affected because query sets leak analogs of validation molecules, so their claims of generalization are not supported.
  • Choosing mean versus median raw EF1% can reverse the apparent ranking of methods; reporting normalized scores, both statistics, and full evaluation code is necessary for fair comparison.
  • Simple duplicate removal cannot salvage the benchmark, so future benchmarks need explicit controls on molecular overlap, scaffold similarity, and query diversity.
  • A no-learnable-parameter memorization baseline can serve as a lower-bound control that any serious virtual-screening benchmark should be checked against.

Reading between the lines

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

  • The same audit recipe could be applied prospectively to any new virtual-screening benchmark: compute a zero-parameter ECFP4 max-similarity baseline as part of the release, and flag any target where its EF1% exceeds a small fraction of the theoretical maximum as failing to test generalization.
  • The metric-sensitivity result suggests that published rankings across virtual-screening methods may be less robust than assumed wherever only one summary statistic is reported; reporting full per-target distributions and normalized scores would settle this for other benchmarks.
  • A testable extension would be to build a leakage-corrected split that removes duplicate inactives and analog pairs at Tanimoto 0.6 and then check whether the memorization baseline's EF1% drops to the level of a random ranker; the paper's claims imply that it would, but the paper does not run this experiment.
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. This manuscript audits the LIT-PCBA benchmark for data leakage and redundancy. Using deterministic RDKit-based comparisons, the authors report 2D-identical molecules shared across training/validation/query splits, thousands of shared inactives, and pervasive analog overlap (e.g., 323 ALDH1 active training–validation pairs at Tanimoto ≥ 0.6). They additionally implement a parameter-free memorization baseline that scores validation molecules by maximum ECFP4 similarity to training actives and report that its median raw EF1% (4.15) equals the median reported for CHEESE's 3D encoders. The paper concludes that LIT-PCBA rewards scaffold memorization rather than generalization and that nearly all published results on it are undermined. Code and data are released.

Significance. If the audit's factual findings hold, it is a valuable cautionary contribution: it adds systematic documentation of split leakage in a widely used benchmark and provides a simple, reproducible baseline demonstrating that training-set similarity alone extracts high enrichment. The deterministic nature of the counts and the release of code are strengths. The significance is reduced, however, by the mismatch between the 'same protocol' parity claim and the actual baseline design, and by the abstract's broader claim about deep-learning and 3D-similarity state-of-the-art results, which is not supported by the comparisons actually reported.

major comments (4)
  1. [Section 3.1 and Tables 2–4] The claim that Mem matches CHEESE's 3D encoders 'under the same multi-ligand protocol' is not supported, because the reference-set protocols differ. Mem scores each validation molecule by maximum ECFP4 similarity to any training active, whereas the CHEESE encoders rank molecules by similarity to the PDB query ligands. The paper's own query-only baseline (Qry in Table 4) has median raw EF1% 1.301, far below the 4.15 median attributed to EspSim and ShapeSim; Mem attains 4.15 only by adding training actives as references. This is exactly the leakage channel documented in Section 5, so the exercise remains a valid leakage demonstration, but it is not an apples-to-apples 'same protocol' comparison. Please revise the wording and present a query-only baseline as the correct protocol-matched comparison.
  2. [Section 3.1.1 and Table 2] The parity conclusion rests on a single statistic, the median raw EF1%, even though the paper itself argues that mean, median, and nEF1% should all be reported. Under the paper's own recommended statistics, Mem does not match the CHEESE encoders: mean raw EF1% is 4.35 for Mem versus 4.66 and 5.21 for CE-E and CE-S; mean nEF1% is 0.063 versus 0.065 and 0.077; median nEF1% is 0.042 versus 0.051 for CE-E. The claim should be restricted to the specific statistic for which it holds and the full metric set should be reported in the main text.
  3. [Abstract and Section 1] The statement that the baseline 'match[es] or exceed[s] ... state-of-the-art deep learning and 3D-similarity models' is not supported by the evidence in the manuscript. Tables 2 and 3 contain no deep-learning baselines; the only head-to-head comparisons are with the two CHEESE 3D encoders. Similarly, the conclusion that 'nearly all published results on LIT-PCBA are undermined' goes beyond the demonstrated scope, which is a leakage demonstration for one family of methods and a set of deterministic overlap counts. Please either provide the missing comparisons or temper the abstract and conclusions to what the data show.
  4. [Section 3.1 and Tables 2–3] The CHEESE per-target values are transcribed from figures of preprint [17] rather than from a machine-readable source. Because the parity claim depends on the exact numbers (e.g., median 4.15), please provide the underlying per-target EF1% values, or the extraction script and figure source, so the comparison is independently checkable from the manuscript or repository.
minor comments (5)
  1. [Section 5.1.2 and Table 6] The text says 'over 350' active training–validation analog pairs, but the listed per-target counts (ALDH1 323, GBA 12, MAPK1 8, plus at least one each in FEN1, PKM2, and IDH1) sum to at least 346; please either report the exact total or align the wording.
  2. [Section 2, paragraph 2] The sentence 'the first step is to extract one or more ligands are from co-crystallized PDB structures' contains a grammatical error; the word 'are' should be removed.
  3. [Section 3.1.1] The term nEF1% is used before it is explicitly defined; please define it at first use in Section 3 (normalization by the target's theoretical maximum enrichment) so that Section 3.1.1 does not introduce it retroactively.
  4. [Section 6.2.1] This section uses EF0.1% while the rest of the paper discusses EF1%; please clarify the relationship between the two metrics and include EF0.1% in the metrics definitions.
  5. [Reference [17]] Reference [17] is given only as a 2024 preprint without a journal, arXiv identifier, or version; please add full bibliographic details so readers can locate the CHEESE paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the memorization baseline is a fixed, parameter-free computation compared against external benchmark values.

full rationale

This paper is an empirical data-integrity audit, not a model-building derivation. Its central quantitative claims are (i) computed overlap counts from fixed RDKit canonicalization and ECFP4 fingerprints with explicit thresholds, and (ii) a zero-parameter memorization baseline (max ECFP4 Tanimoto to training actives) whose outputs are compared with EF1% values transcribed from CHEESE Figs. 6-7. Nothing in the baseline is fitted to the CHEESE numbers; the thresholds (Tc >= 0.6, 0.85; MCS >= 0.9) are stated a priori, and the leakage findings follow from direct set operations on the released splits. The paper cites external work and acknowledges a protocol clarification from a CHEESE author, but that is not a load-bearing self-citation: the comparison could stand or fall on the transcribed numbers and the public code release. The skeptic concerns about Mem versus CHEESE reference-set protocol and the abstract's 'deep learning' phrasing concern the fairness and scope of the comparison, not circularity; likewise the absence of machine-readable CHEESE values is a reproducibility limitation. No step derives its conclusion from the quantity it purports to test, so the circularity score is 0.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The audit introduces no fitted parameters and postulates no new entities. Its inputs are similarity thresholds chosen by hand and a stereo-agnostic identity definition, plus external CHEESE numbers transcribed from figures. None of these are derived from the benchmark outcomes being criticized, so the circularity burden is minimal.

free parameters (2)
  • Inter-set analog Tanimoto threshold = 0.6
    Chosen in Section 5.1.2 as a 'conservative' cutoff to define analog pairs; the reported counts (323 ALDH1 pairs, etc.) would change with a different threshold.
  • Intra-set analog thresholds = Tc >= 0.85 or MCS >= 0.9
    Used in Table 6 for intra-set redundancy; near-duplicate cutoffs chosen by the authors, not derived from data.
assumptions (3)
  • domain assumption Overlap between training and validation molecules, at the detected levels, inflates EF1% and AUROC by rewarding memorization.
    This is the interpretive premise of Sections 4-6; the paper demonstrates it only for the memorization baseline on EF1%, not for all model classes or AUROC.
  • domain assumption The CHEESE EF1% values in Tables 2 and 3 were transcribed accurately from Figs. 6 and 7 of the CHEESE preprint [17].
    The parity claim depends on these external numbers, and the paper provides no machine-readable source.
  • domain assumption RDKit canonical SMILES after stereochemistry removal is a valid definition of 2D-identity for leakage counting.
    Standard cheminformatics practice, but the specific counting results depend on this choice, as the paper acknowledges in Section 4.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Data Leakage and Redundancy in the LIT-PCBA Benchmark." pith.science (2026). https://pith.science/paper/23HOTS6V

@misc{pith2026250721404,
  author       = {Pith},
  title        = {Pith review of: Data Leakage and Redundancy in the LIT-PCBA Benchmark},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/23HOTS6V}},
  note         = {Machine review of arXiv:2507.21404}
}
abstract

LIT-PCBA is widely used to benchmark virtual screening models, but our audit reveals that it is fundamentally compromised. We find extensive data leakage and molecular redundancy across its splits, including 2D-identical ligands within and across partitions, pervasive analog overlap, and low-diversity query sets. In ALDH1 alone, for instance, 323 active training -- validation analog pairs occur at ECFP4 Tanimoto similarity $\geq 0.6$; across all targets, 2,491 2D-identical inactives appear in both training and validation, with very few corresponding actives. These overlaps allow models to succeed through scaffold memorization rather than generalization, inflating enrichment factors and AUROC scores. These flaws are not incidental -- they are so severe that a trivial memorization-based baseline with no learnable parameters can exploit them to match or exceed the reported performance of state-of-the-art deep learning and 3D-similarity models. As a result, nearly all published results on LIT-PCBA are undermined. Even models evaluated in "zero-shot" mode are affected by analog leakage into the query set, weakening claims of generalization. In its current form, the benchmark does not measure a model's ability to recover novel chemotypes and should not be taken as evidence of methodological progress. All code, data, and baseline implementations are available at: https://github.com/sievestack/LIT-PCBA-audit

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 32 canonical work pages

  1. [17]

    Cheese: 3d shape and electrostatic virtual screening in a vector space

    Miroslav Lžičař and Hamza Gamouh. Cheese: 3d shape and electrostatic virtual screening in a vector space. 2024

    work page 2024

  2. [1]

    Benchmarking data sets for the evaluation of virtual ligand screening methods: review and perspectives.Journal of chemical information and modeling, 55(7):1297–1307, 2015

    Nathalie Lagarde, Jean-François Zagury, and Matthieu Montes. Benchmarking data sets for the evaluation of virtual ligand screening methods: review and perspectives.Journal of chemical information and modeling, 55(7):1297–1307, 2015

    work page 2015

  3. [2]

    Benchmarking methods and data sets for ligand enrichment assessment in virtual screening

    Jie Xia, Ermias Lemma Tilahun, Terry-Elinor Reid, Liangren Zhang, and Xiang Simon Wang. Benchmarking methods and data sets for ligand enrichment assessment in virtual screening. Methods, 71:146–157, 2015. 16

    work page 2015

  4. [3]

    Community benchmarks for virtual screening.Journal of computer-aided molecular design, 22(3):193–199, 2008

    John J Irwin. Community benchmarks for virtual screening.Journal of computer-aided molecular design, 22(3):193–199, 2008

    work page 2008

  5. [4]

    Lit-pcba: an unbiased data set for machine learning and virtual screening.Journal of chemical information and modeling, 60(9):4263–4273, 2020

    Viet-Khoa Tran-Nguyen, Célien Jacquemard, and Didier Rognan. Lit-pcba: an unbiased data set for machine learning and virtual screening.Journal of chemical information and modeling, 60(9):4263–4273, 2020

    work page 2020

  6. [5]

    Hidden bias in the dud-e dataset leads to misleading performance of deep learning in structure-based virtual screening.PloS one, 14(8):e0220113, 2019

    Lieyang Chen, Anthony Cruz, Steven Ramsey, Callum J Dickson, Jose S Duca, Viktor Hornak, David R Koes, and Tom Kurtzman. Hidden bias in the dud-e dataset leads to misleading performance of deep learning in structure-based virtual screening.PloS one, 14(8):e0220113, 2019

    work page 2019

  7. [6]

    Directory of useful decoys, enhanced (dud-e): better ligands and decoys for better benchmarking.Journal of medicinal chemistry, 55(14):6582–6594, 2012

    Michael M Mysinger, Michael Carchia, John J Irwin, and Brian K Shoichet. Directory of useful decoys, enhanced (dud-e): better ligands and decoys for better benchmarking.Journal of medicinal chemistry, 55(14):6582–6594, 2012

    work page 2012

  8. [7]

    Maximum unbiased validation (muv) data sets for virtual screening based on pubchem bioactivity data.Journal of chemical information and modeling, 49(2):169–184, 2009

    Sebastian G Rohrer and Knut Baumann. Maximum unbiased validation (muv) data sets for virtual screening based on pubchem bioactivity data.Journal of chemical information and modeling, 49(2):169–184, 2009

    work page 2009

Show all 32 references
  1. [8]

    Recognizing pitfalls in virtual screening: a critical review.Journal of chemical information and modeling, 52(4):867–881, 2012

    Thomas Scior, Andreas Bender, Gary Tresadern, José L Medina-Franco, Karina Martínez- Mayorga, Thierry Langer, Karina Cuanalo-Contreras, and Dimitris K Agrafiotis. Recognizing pitfalls in virtual screening: a critical review.Journal of chemical information and modeling, 52(4):8...

    2012

  2. [9]

    Recent progress on the prospective application of machine learning to structure-based virtual screening.Current opinion in chemical biology, 65:28–34, 2021

    Ghita Ghislat, Taufiq Rahman, and Pedro J Ballester. Recent progress on the prospective application of machine learning to structure-based virtual screening.Current opinion in chemical biology, 65:28–34, 2021

    2021

  3. [10]

    Deep learning in virtual screening: recent applications and developments.International journal of molecular sciences, 22(9):4435, 2021

    Talia B Kimber, Yonghui Chen, and Andrea Volkamer. Deep learning in virtual screening: recent applications and developments.International journal of molecular sciences, 22(9):4435, 2021

    2021

  4. [11]

    Artificial intelligence in virtual screening: Models versus experiments.Drug Discovery Today, 27(7):1913– 1923, 2022

    N Arul Murugan, Gnana Ruba Priya, G Narahari Sastry, and Stefano Markidis. Artificial intelligence in virtual screening: Models versus experiments.Drug Discovery Today, 27(7):1913– 1923, 2022

    1913

  5. [12]

    Recent advances in ai- driven protein-ligand interaction predictions.Current Opinion in Structural Biology, 92:103020, 2025

    Jaemin Sim, Dongwoo Kim, Bomin Kim, Jieun Choi, and Juyong Lee. Recent advances in ai- driven protein-ligand interaction predictions.Current Opinion in Structural Biology, 92:103020, 2025

    2025

  6. [13]

    Recent developments in ultralarge and structure-based virtual screening approaches

    Christoph Gorgulla. Recent developments in ultralarge and structure-based virtual screening approaches. Annual Review of Biomedical Data Science, 6(1):229–258, 2023

    2023

  7. [14]

    Scaling structure aware virtual screening to billions of molecules with sprint.ArXiv, pages arXiv–2411, 2025

    Andrew T McNutt, Abhinav K Adduri, Caleb N Ellington, Monica T Dayao, Eric P Xing, Hosein Mohimani, and David R Koes. Scaling structure aware virtual screening to billions of molecules with sprint.ArXiv, pages arXiv–2411, 2025

    2025

  8. [15]

    Drugclip: Contrastive protein-molecule representation learning for virtual screening

    Bowen Gao, Bo Qiang, Haichuan Tan, Yinjun Jia, Minsi Ren, Minsi Lu, Jingjing Liu, Wei-Ying Ma, and Yanyan Lan. Drugclip: Contrastive protein-molecule representation learning for virtual screening. Advances in Neural Information Processing Systems, 36:44595–44614, 2023

    2023

  9. [16]

    Agamemnon Krasoulis, Nick Antonopoulos, Vassilis Pitsikalis, and Stavros Theodorakis. Denvis: scalable and high-throughput virtual screening using graph neural networks with atomic and 17 surface protein pocket features.Journal of Chemical Information and Modeling, 62(19):4642...

    2022

  10. [18]

    Milcdock: machine learning enhanced consensus docking for virtual screening in drug discovery

    Connor J Morris, Jacob A Stern, Brenden Stark, Max Christopherson, and Dennis Della Corte. Milcdock: machine learning enhanced consensus docking for virtual screening in drug discovery. Journal of Chemical Information and Modeling, 62(22):5342–5350, 2022

    2022

  11. [19]

    A. T. McNutt, P. G. Francoeur, R. Aggarwal, T. Masuda, R. Meli, M. Ragoza, J. Sunseri, and D. R. Koes. Gnina 1.0: Molecular docking with deep learning.Journal of Cheminformatics, 13(1):43, 2021

    2021

  12. [20]

    Improved deepdta for protein–ligand binding affinity prediction with transformer and contrastive learning.Bioinformatics, 40(1):btad683, 2024

    Xinqiao Li, Xiaoxia Du, Yakun Gao, Hongri Zhang, Pengfei Yu, Hai Wang, Jian Li, Min Chen, and Lidian Huang. Improved deepdta for protein–ligand binding affinity prediction with transformer and contrastive learning.Bioinformatics, 40(1):btad683, 2024

    2024

  13. [21]

    J. P. Janet, T. Z. H. Gani, S. M. Kearnes, et al. Latent biases in machine learning models for predicting binding affinities.Journal of Chemical Information and Modeling, 63(4):984–1000, 2023

    2023

  14. [22]

    Latent biases in machine learning models for predicting binding affinities using popular data sets.ACS omega, 8(2):2389–2397, 2023

    Ganesh Chandan Kanakala, Rishal Aggarwal, Divya Nayar, and U Deva Priyakumar. Latent biases in machine learning models for predicting binding affinities using popular data sets.ACS omega, 8(2):2389–2397, 2023

    2023

  15. [23]

    Virtual screening: an endless staircase?Nature Reviews Drug Discovery, 9(4):273–276, 2010

    Gisbert Schneider. Virtual screening: an endless staircase?Nature Reviews Drug Discovery, 9(4):273–276, 2010

    2010

  16. [24]

    Property-unmatched decoys in docking benchmarks

    Reed M Stein, Ying Yang, Trent E Balius, Matt J O’Meara, Jiankun Lyu, Jennifer Young, Khanh Tang, Brian K Shoichet, and John J Irwin. Property-unmatched decoys in docking benchmarks. Journal of chemical information and modeling, 61(2):699–714, 2021

    2021

  17. [25]

    Most ligand-based classification benchmarks reward memorization rather than generalization

    Izhar Wallach and Abraham Heifets. Most ligand-based classification benchmarks reward memorization rather than generalization. Journal of chemical information and modeling, 58(5):916–932, 2018

    2018

  18. [26]

    A comprehensive comparative assessment of 3d molecular similarity tools in ligand-based virtual screening.Briefings in bioinformatics, 22(6), 2021

    Zhenla Jiang, Jianrong Xu, Aixia Yan, and Ling Wang. A comprehensive comparative assessment of 3d molecular similarity tools in ligand-based virtual screening.Briefings in bioinformatics, 22(6), 2021

    2021

  19. [27]

    Conformational space profiling enhances generic molecular representation for ai-powered ligand-based drug discovery.Advanced Science, 11(40):2403998, 2024

    Lin Wang, Shihang Wang, Hao Yang, Shiwei Li, Xinyu Wang, Yongqi Zhou, Siyuan Tian, Lu Liu, and Fang Bai. Conformational space profiling enhances generic molecular representation for ai-powered ligand-based drug discovery.Advanced Science, 11(40):2403998, 2024

    2024

  20. [28]

    Protein–ligand docking in the machine- learning era

    Chao Yang, Eric Anthony Chen, and Yingkai Zhang. Protein–ligand docking in the machine- learning era. Molecules, 27(14):4568, 2022

    2022

  21. [29]

    Delta machine learning to improve scoring-ranking-screening performances of protein–ligand scoring functions.Journal of chemical information and modeling, 62(11):2696–2712, 2022

    Chao Yang and Yingkai Zhang. Delta machine learning to improve scoring-ranking-screening performances of protein–ligand scoring functions.Journal of chemical information and modeling, 62(11):2696–2712, 2022

    2022

  22. [30]

    Rdkit: Open-source cheminformatics.https://www.rdkit.org

    RDKit Development Team. Rdkit: Open-source cheminformatics.https://www.rdkit.org. Accessed: 2025-08-03. 18

    2025

  23. [31]

    Alphafold2 structures guide prospective ligand discovery.Science, 384(6702):eadn6354, 2024

    Jiankun Lyu, Nicholas Kapolka, Ryan Gumpper, Assaf Alon, Liang Wang, Manish K Jain, Ximena Barros-Álvarez, Kensuke Sakamoto, Yoojoong Kim, Jeffrey DiBerto, et al. Alphafold2 structures guide prospective ligand discovery.Science, 384(6702):eadn6354, 2024

    2024

  24. [32]

    Matthias R Bauer, Tamer M Ibrahim, Simon M Vogel, and Frank M Boeckler. Evaluation and optimization of virtual screening workflows with dekois 2.0–a public library of challenging docking benchmark sets.Journal of chemical information and modeling, 53(6):1447–1462, 2013. 19

    2013

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.