A Machine Learning Benchmarking Framework for Lipid Nanoparticle Transfection Efficiency Prediction

Asal Mehradfar; Glen S. Kwon; Jose Miguel Hernandez-Lobato; Mahdi Soltanolkotabi; Mohammad Shahab Sepehri; Morteza Rasoulianboroujeni; Salman Avestimehr

REVIEW 4 major objections 5 minor 63 references

A Machine Learning Benchmarking Framework for Lipid Nanoparticle Transfection Efficiency Prediction

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 This paper claims that explicit molecular substructure encoding, especially count-based Morgan fingerprints combined with expert descriptors, outperforms graph-based deep learning models for predicting lipid nanoparticle transfection…

desk verdict Corrected dataset and a plausible negative result on AGILE, but the single-split evaluation with test-set feature selection leaves the headline ranking unproven. read the letter →

arxiv 2507.03209 v2 pith:R4LEQ6RU submitted 2025-07-03 q-bio.QM cs.CEcs.LGq-bio.MN

Asal Mehradfar , Mohammad Shahab Sepehri , Jose Miguel Hernandez-Lobato , Glen S. Kwon , Mahdi Soltanolkotabi , Salman Avestimehr , Morteza Rasoulianboroujeni This is my paper · ORCID

classification q-bio.QMcs.CEcs.LGq-bio.MN

keywords lipidnanoparticlestransfectionefficiencypredictionmachinelearningbenchmarkingMorganfingerprintsmolecularrepresentationgraphneuralnetworksionizablelipidsvirtualscreening

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 introduces LANTERN, a benchmarking framework that predicts how well an ionizable lipid will deliver RNA by comparing many molecular representations and model architectures on one curated dataset. Its central claim is that explicit substructure encoding—count-based Morgan fingerprints, optionally combined with 210 expert RDKit descriptors—consistently outperforms graph-based models that learn their own embeddings, including the current AGILE model. On the refined 1,100-lipid HeLa dataset, a multilayer perceptron trained on Morgan fingerprints plus expert descriptors reaches $R^2 = 0.8161$ with Pearson $r = 0.9053$, while AGILE reaches $R^2 = 0.2655$, $r = 0.5488$ under the same random split. The paper also reports correcting 235 label mismatches and removing 100 cis/trans duplicate SMILES from the original AGILE data, and argues that AGILE's published accuracy was inflated by including training and validation data in its evaluation.

What carries the argument

The load-bearing object is the count-based Morgan fingerprint: a 2048-dimensional vector in which each entry records how many times a given circular substructure appears around an atom, computed with a standard cheminformatics library. The paper uses it both alone and combined with 210 expert-derived physicochemical descriptors such as molecular weight, logP, and topological polar surface area. This fingerprint does the work because it gives small datasets enough explicit chemistry for simple models to learn from, and because adding it to AGILE's own learned features lifts AGILE from $R^2 = 0.2655$ to $0.706$. The contrasting objects are learned embeddings from graph neural networks and graph transformers, which the paper shows are less predictive unless augmented with the same explicit substructure features.

What would settle it

Re-run the entire benchmark on an independent transfection dataset (for example the RAW264.7 measurements the AGILE study also reports, or a newly synthesized lipid library) with the same eight models and feature sets; if Morgan-fingerprint MLP does not maintain its $R^2$ and ranking advantage there, the paper's central claim would be contradicted. A cheaper check is to recompute Table 2 with the corrected labels only and no duplicate removal, to see whether AGILE's gap to MLP shrinks materially, which would locate the cause in the cis/trans curation step.

Watch

Extended reading notes

Core claim

The paper's central claim is that, for LNP transfection efficiency, a molecular representation that explicitly enumerates substructures carries more predictive signal than learned graph embeddings. The authors support this by benchmarking eight architectures—MLP, transformer, SVR, random forest, kNN, KPGT, Chemprop, and AGILE—on one curated HeLa dataset derived from the AGILE platform. The best model, a seven-layer MLP using count-based Morgan fingerprints plus expert descriptors, achieves $R^2 = 0.8161$ and $r = 0.9053$ on the test split; AGILE, run from its official implementation, achieves $R^2 = 0.2655$ and $r = 0.5488$. The same ordering holds under Murcko scaffold splitting, where every model beats AGILE ($R^2 = 0.0057$) and kNN becomes the best performer. The paper further claims that AGILE's originally published accuracy was inflated by computing confusion matrices on training, validation, and test data together, and that once all models are evaluated strictly on a test split, even kNN and a transformer outperform AGILE.

Load-bearing premise

The benchmark's conclusions depend on treating the original experimental source data as ground truth and on the manual corrections—235 label re-mappings and removal of 100 cis/trans duplicates—not having introduced systematic bias that favors explicit fingerprint features over graph embeddings.

Editorial extensions

If this is right

Any future LNP transfection benchmark should include count-based Morgan fingerprints alone and with expert descriptors as mandatory baselines; the paper shows these features are the strongest predictors across five non-graph architectures.
Graph-based models cannot be assumed superior on small molecular datasets: on this data, kNN, random forest, SVR, transformer, and MLP all beat AGILE under random splitting, and several beat the other graph baselines.
AGILE's reported ranking accuracy is not comparable to later models unless recomputed on the test split only; the paper's confusion-matrix analysis puts AGILE's high-transfection bin accuracy at 26.3% versus 73.7% for the best MLP.
The released curated dataset of 1,100 unique lipids gives the field a common testbed; models trained on the original 1,200-point set with duplicate and mislabeled SMILES may be learning inconsistent targets.
For practical screening, the best LANTERN model identifies top-percentile lipids at 73.7% accuracy, meaning a larger fraction of synthesized candidates should be active, cutting the experimental cost of lead discovery.

Reading between the lines

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

If the dataset corrections are right, earlier comparisons made on the unrefined AGILE data should be re-examined; some of AGILE's reported edge may come from the mislabeled and duplicate SMILES rather than from the model architecture.
A natural extension is to run the same benchmark on AGILE's RAW264.7 measurements: if Morgan-plus-MLP does not retain its advantage on that second cell line, the claim is specific to HeLa rather than general.
The pattern suggests a broader design rule for low-data molecular property prediction: choose features that inject known chemistry before choosing architecture, and treat graph embeddings as needing either much larger pretraining or explicit feature augmentation.
One testable prediction follows from the paper's mechanism: adding count-based Morgan fingerprints to KPGT or Chemprop should narrow the gap to MLP more than adding generic expert descriptors, mirroring what the paper observes for AGILE.

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.

Desk Editor's Note

read the letter

Colleague,

The useful thing here is the corrected dataset and the reminder that simple fingerprints are a strong baseline. The weakness is not the curation; it's the single-split evaluation and test-set feature selection.

What's genuinely new: they audited the AGILE dataset and found 235 label mismatches and 100 duplicate SMILES (cis/trans conflated), and they release the refined 1,100-lipid set. That's a real service to the LNP-ML community. The model comparison is broad—MLP, SVR, RF, kNN, transformer, Chemprop, KPGT, AGILE—across three feature sets, and the finding that an MLP on count-based Morgan + Expert descriptors beats AGILE (R2 0.816 vs 0.266) is plausible and consistent with prior work showing GNNs struggle on small molecular datasets. Including a Murcko scaffold split is also good; it honestly shows the ranking is split-dependent (kNN becomes best), which complicates the headline claim.

The soft spots are in the evaluation protocol. All metrics come from a single random split with no error bars, no repeated seeds, no significance tests. They selected the best feature set per model using test-set performance, which risks overfitting the comparison. AGILE is not re-tuned, so the comparison is asymmetric. The stress-test is right that the curation is a secondary issue: AGILE's test accuracy changes from 24.17% to 24.55% after refinement, so the gap isn't explained by the curation. But we can't tell how much of the gap is genuine predictive superiority versus split noise or selection bias. The Murcko split result alone shows the ranking is unstable. A minor point: the relative error metric divides by the true value, which misbehaves near zero.

Who's this for? Anyone building or evaluating LNP transfection predictors. The dataset and the negative result on AGILE are valuable even if the benchmark needs hardening. It deserves a serious referee, but the referee should push for repeated nested splits, confidence intervals, and a fairer AGILE baseline before the "essential baselines" claim is accepted.

You should send this to review, with the expectation of a major revision.

Referee Report

4 major / 5 minor

Summary. The paper introduces LANTERN, a benchmarking framework for predicting lipid nanoparticle (LNP) transfection efficiency from ionizable lipid structures. The authors curate a refined dataset of 1,100 lipids from the HeLa transfection data of Xu et al. by correcting 235 label mismatches and removing 100 duplicate SMILES entries arising from cis/trans isomerism. They then benchmark a range of models—MLP, transformer, SVR, RF, kNN, KPGT, Chemprop, and AGILE—across multiple molecular representations (count-based Morgan fingerprints, Expert RDKit descriptors, Grover embeddings) under random and Murcko scaffold splits. The central claim is that explicit molecular substructure encoding, particularly count-based Morgan fingerprints combined with Expert descriptors and an MLP, consistently achieves the highest predictive accuracy (R2 = 0.8161, r = 0.9053), substantially outperforming AGILE (R2 = 0.2655, r = 0.5488) on the same dataset.

Significance. If the results are robust, the paper provides a useful community resource: a carefully curated dataset, an open-source benchmarking library, and a broad comparison of representation–architecture combinations. The inclusion of a scaffold-split analysis and ranking-based evaluation is valuable, and the public release of code and data supports reproducibility. The comparison against AGILE, the current domain-specific state of the art, is important for the field. However, the strength of the headline claims is currently limited by the evaluation protocol: single-split metrics, test-set-based feature selection, and lack of uncertainty quantification mean that the reported rankings may not be stable. The scaffold-split results themselves show a different top model (kNN versus MLP), undermining the 'consistently' language. These issues are addressable with additional experiments.

major comments (4)

[Section 2.4, Table 2] All reported regression metrics come from a single random split with no repeated runs, confidence intervals, or significance tests. With only 1,100 molecules and high-dimensional features (2048 Morgan + 210 RDKit descriptors), the observed gaps (e.g., MLP R2 = 0.8161 vs. KPGT R2 = 0.6655 vs. SVR R2 = 0.7285) may reflect split noise rather than genuine differences. The claim that models using explicit substructure encoding 'consistently achieve the highest predictive accuracy' requires either repeated nested cross-validation, bootstrap confidence intervals, or paired significance tests across multiple splits.
[Section 4.1, Tables 1–3] The best feature set for each model appears to be selected using the same test split that is subsequently used for reporting performance (e.g., Figure 4 states each model is evaluated 'using its optimal feature representation, as determined by previous analyses'). This test-set-based feature selection introduces optimism into the reported metrics and creates an unfair advantage for models with more feature-set variants. A separate validation set or nested cross-validation should be used for feature and hyperparameter selection before computing test metrics.
[Section 2.5, Table 3] Under the Murcko scaffold split, the ranking changes substantially: kNN achieves the best test R2 (0.6146 with Morgan features) while MLP with Morgan+Expert drops to R2 = 0.4532, which is lower than Chemprop's R2 = 0.5129 and KPGT's R2 = 0.4878. This directly contradicts the abstract's claim of consistent superiority of explicit-substructure models. The paper should either restrict the headline claim to the random-split setting or explicitly discuss the split-dependence as a limitation of the central conclusion.
[Section 4.1] The paper states that 'all model hyperparameters were tuned to ensure optimal performance, except for AGILE, which was used directly from the original implementation without any modifications.' Since AGILE is the primary baseline and the target of the comparison, it is important to report whether AGILE's hyperparameters (e.g., learning rate, number of epochs, early stopping criteria) were re-tuned on the curated dataset. If not, the observed large gap between AGILE and other models may partly reflect an un-tuned baseline. At minimum, a tuned AGILE variant should be included to make the comparison equitable.

minor comments (5)

[Title] The arXiv metadata title reads 'A Machine Learning Benchmarking Framework for Lipid Nanoparticle Transfection Efficiency Prediction,' while the manuscript title is 'LANTERN: A Machine Learning Framework for Lipid Nanoparticle Transfection Efficiency Prediction.' Please align the two titles.
[Section 2.2] The curation process assumes the experimental 'source data' file is ground truth for correcting the 235 label mismatches. The paper would benefit from an explicit statement of any validation performed and from releasing the mapping/correction tables to allow others to audit the curation.
[Section 2.6, Figure 5] The relative error histograms are described qualitatively (e.g., 'narrower and more concentrated' for MLP). Quantitative summaries such as median, interquartile range, or the fraction of predictions within a given error threshold would make the comparison more precise.
[Section 2.6, Figure 6] The ranking-based accuracies reported in the text (e.g., MLP average 50.0%, high bins 73.7%) should state the number of test samples, the bin definitions, and whether these numbers are averages over the six percentile bins or a single aggregate accuracy.
[Table 1] For the AGILE rows in Table 1, it would be helpful to note explicitly that AGILE always includes its end-to-end learned representation in addition to the listed feature sets, since the table otherwise reads as if the listed features are the only inputs.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: LANTERN is an empirical benchmark on an external dataset, and the reported rankings do not reduce to any fitted input or load-bearing self-citation.

full rationale

LANTERN is an empirical benchmarking study, not a derivation. The central claim—that explicit substructure encodings, particularly count-based Morgan fingerprints combined with Expert descriptors and an MLP, outperform graph-based models—is established by fitting independent models to an external dataset (Xu et al.'s HeLa transfection data) and comparing held-out test metrics. None of the reported R2, RMSE, MAE, or r values is obtained by algebraic manipulation of the feature definitions; each is a fitted model evaluation on a test split. The AGILE baseline is run from its official implementation, and the comparison does not rely on any parameter fitted to the target conclusion. The dataset curation steps (correcting 235 label mismatches and removing 100 duplicate cis/trans SMILES entries) are input-processing choices; even if contestable, they do not define the benchmark outcome by construction. The only self-citation is reference [60], used to support the generic statement that ML models are widely applied across domains, and it is not load-bearing. There is no imported uniqueness theorem, no ansatz smuggled in via citation, and no renamed known result. Methodological concerns such as single-split evaluation and selecting feature representations based on test performance are validity risks, not circularity: the predictions are not equal-by-construction to the inputs. The paper's derivation chain is therefore self-contained with respect to circularity.

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

This is an empirical benchmarking paper, not a theory derivation. It introduces no free parameters in the scientific-law sense and no new postulated entities. Its conclusions rest on data curation assumptions, a single data split, and standard representation tooling, which is why the axioms above are the load-bearing supports.

assumptions (3)

domain assumption The experimental 'source data' file from Xu et al. is the true ground truth for transfection values, and the 235 corrected labels and 100 duplicate removals are accurate.
Section 2.2 states that corrections were made manually assuming the source data as ground truth. If this mapping is wrong, all subsequent comparisons and rankings are invalid.
domain assumption A single random split of the refined dataset yields a representative test distribution for ranking models.
Sections 2.3 and 2.4 use one random split to report all R2, RMSE, MAE, and r values. Without repeated splits or cross-validation, the ranking could be sensitive to that particular split.
domain assumption Standard cheminformatics tooling (DeepChem, RDKit) correctly computes Morgan fingerprints and Expert descriptors for ionizable lipid structures.
Section 4.2 relies on default implementations from DeepChem and RDKit. Bugs or version-specific behavior in these libraries would affect the reported results, though this is a common and reasonable assumption.

how reviews work

0 comments

Cite this review

Pith. "Pith review of A Machine Learning Benchmarking Framework for Lipid Nanoparticle Transfection Efficiency Prediction." pith.science (2026). https://pith.science/paper/R4LEQ6RU

@misc{pith2026250703209,
  author       = {Pith},
  title        = {Pith review of: A Machine Learning Benchmarking Framework for Lipid Nanoparticle Transfection Efficiency Prediction},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R4LEQ6RU}},
  note         = {Machine review of arXiv:2507.03209}
}

read the original abstract

The discovery of new ionizable lipids for efficient lipid nanoparticle (LNP)-mediated RNA delivery remains a major bottleneck in RNA therapeutics development. Recent advances demonstrate the potential of machine learning (ML) models to predict transfection efficiency directly from lipid structure, enabling high-throughput virtual screening and accelerating lead identification. However, as new models for LNP transfection prediction continue to emerge, the lack of rigorous and standardized benchmarking poses a significant risk and may undermine confidence in their reliability for discovery. Here, we present a robust ML benchmarking framework for evaluating transfection prediction models based on ionizable lipid structures. This framework systematically benchmarks diverse molecular representations paired with a broad range of ML architectures spanning traditional models, feedforward neural networks, and state-of-the-art graph-based methods. In addition, the presented framework supports assessment of model generalization and evaluates prediction reliability beyond standard regression metrics. Using a curated dataset of 1,100 unique ionizable lipid structures derived from the HeLa transfection dataset originally reported by Xu et al., we show that within this framework, models leveraging explicit molecular substructure encoding consistently achieve the highest predictive accuracy and should serve as essential baselines for the development of new, more sophisticated models. In contrast, some current graph-based models, including AGILE, Chemprop, and KPGT, tend to show comparatively lower accuracy. The presented framework provides a standardized, transparent, and comprehensive benchmarking resource that enables meaningful comparison of emerging architectures and establishes strong baselines for future development of predictive models in lipid-based RNA delivery.

Figures

Figures reproduced from arXiv: 2507.03209 by the authors.

**Figure 1.** Overview of molecular representation techniques. (a) Morgan fingerprints encode molecular structures as binary vectors based on circular substructures. (b) Expert descriptors use predefined molecular properties, such as hydrogen bonding capacity, charge, and mass. (c) Graph transformers (e.g., Grover, KPGT) capture structural relationships through self-attention mechanisms, generating molecular embeddings. (d) Graph… view at source ↗

**Figure 2.** t-SNE visualization of feature representations. Each plot illustrates the distribution of LNP representations in a two-dimensional space, colored by transfection efficiency values. (a) Morgan fingerprints form well-defined clusters, suggesting they effectively capture key molecular substructures, leading to superior predictive performance. (b) Expert descriptors exhibit a more continuous spread, making them less dis… view at source ↗

**Figure 3.** Prediction vs. true value plots for different feature sets. (a) AGILE model using end-to-end features. (b–h) MLP model trained with different feature sets, including Morgan, Expert, and Grover representations individually and in combination. Morgan fingerprints combined with Expert descriptors achieve the best performance with the lowest RMSE, demonstrating the advantage of combining domain-specific features with su… view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Prediction vs. true value plots for different models. Each model is evaluated using its optimal feature representation, as determined by previous analyses. (a–h) show the predictive performance of MLP, Transformer, SVR, RF, kNN, KPGT, Chemprop, and AGILE, respectively.…

**Figure 5.** Figure 5: Relative error distribution across different models. This figure presents the relative error histograms for nine models evaluated, including two MLP variants trained with Morgan fingerprints (MLP-M) and Morgan fingerprints combined with expert descriptors (MLP-M-E). Fo…

**Figure 6.** Figure 6: Confusion matrices for different models using the AGILE ranking-based evaluation approach. Each matrix represents the performance of a model on the test set, where the best-performing feature set is selected for each model. The confusion matrices are constructed based …

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 57 canonical work pages

[1]

Non-viral vectors for rna delivery

Yan, Y .et al. Non-viral vectors for rna delivery. Journal of Controlled Release 342, 241–279 (2022)

work page 2022
[2]

& Jin, H

Zhu, Y ., Zhu, L., Wang, X. & Jin, H. Rna-based therapeutics: an overview and prospectus.Cell death & disease 13, 644 (2022)

work page 2022
[3]

L., Ciccone, G., Esposito, C

Ibba, M. L., Ciccone, G., Esposito, C. L., Catuogno, S. & Giangrande, P. H. Advances in mrna non-viral delivery approaches. Advanced drug delivery reviews 177, 113930 (2021)

work page 2021
[4]

Kim, M. et al. Engineered ionizable lipid nanoparticles for targeted delivery of rna therapeutics into different types of cells in the liver. Science advances 7, eabf4398 (2021)

work page 2021
[5]

Wittrup, A. et al. Visualizing lipid-formulated sirna release from endosomes and target gene knockdown. Nature biotechnology 33, 870–876 (2015)

work page 2015
[6]

Xu, E., Saltzman, W. M. & Piotrowski-Daspit, A. S. Escaping the endosome: assessing cellular trafficking mechanisms of non-viral vehicles. Journal of Controlled Release 335, 465–480 (2021)

work page 2021
[7]

Albertsen, C. H. et al. The role of lipid components in lipid nanoparticles for vaccines and gene therapy. Advanced drug delivery reviews 188, 114416 (2022)

work page 2022
[8]

Cheng, Q. et al. Selective organ targeting (sort) nanoparticles for tissue-specific mrna delivery and crispr–cas gene editing. Nature nanotechnology 15, 313–320 (2020)

work page 2020

Show all 63 references

[9]

Ni, H. et al. Piperazine-derived lipid nanoparticles deliver mrna to immune cells in vivo. Nature Communications 13, 4766 (2022)

2022
[10]

Gan, Z. et al. Nanoparticles containing constrained phospholipids deliver mrna to liver immune cells in vivo without targeting ligands. Bioengineering & translational medicine 5, e10161 (2020)

2020
[11]

Qiu, M. et al. Lung-selective mrna delivery of synthetic lipid nanoparticles for the treatment of pulmonary lymphangioleiomy- omatosis. Proceedings of the national academy of sciences 119, e2116271119 (2022)

2022
[12]

Multi-step screening of dna/lipid nanoparticles and co-delivery with sirna to enhance and prolong gene expression

Zhu, Y .et al. Multi-step screening of dna/lipid nanoparticles and co-delivery with sirna to enhance and prolong gene expression. Nature communications 13, 4282 (2022)

2022
[13]

Li, B. et al. Combinatorial design of nanoparticles for pulmonary mrna delivery and genome editing. Nature biotechnology 41, 1410–1415 (2023)

2023
[14]

Liu, S. et al. Membrane-destabilizing ionizable phospholipids for organ-selective mrna delivery and crispr–cas gene editing. Nature materials 20, 701–710 (2021)

2021
[15]

Liu, S. et al. Zwitterionic phospholipidation of cationic polymers facilitates systemic mrna delivery to spleen and lymph nodes. Journal of the American Chemical Society 143, 21321–21330 (2021)

2021
[16]

Radmand, A. et al. The transcriptional response to lung-targeting lipid nanoparticles in vivo. Nano letters 23, 993–1002 (2023)

2023
[17]

Radmand, A. et al. Cationic cholesterol-dependent lnp delivery to lung stem cells, the liver, and heart. Proceedings of the National Academy of Sciences 121, e2307801120 (2024)

2024
[18]

Miao, L. et al. Delivery of mrna vaccines with heterocyclic lipids increases anti-tumor efficacy by sting-mediated immune cell activation. Nature biotechnology 37, 1174–1185 (2019)

2019
[19]

Li, B. et al. Accelerating ionizable lipid discovery for mrna delivery using machine learning and combinatorial chemistry. Nature Materials 1–7 (2024)

2024
[20]

Han, X. et al. An ionizable lipid toolbox for rna delivery. Nature communications 12, 7233 (2021)

2021
[21]

Y ., Zhang, Y ., Jia, Y

Ding, D. Y ., Zhang, Y ., Jia, Y . & Sun, J. Machine learning-guided lipid nanoparticle design for mrna delivery.arXiv preprint arXiv:2308.01402 (2023)

2023 arXiv
[22]

Moayedpour, S. et al. Representations of lipid nanoparticles using large language models for transfection efficiency prediction. Bioinformatics btae342 (2024)

2024
[23]

Self-supervised graph transformer on large-scale molecular data

Rong, Y .et al. Self-supervised graph transformer on large-scale molecular data. Advances in neural information processing systems 33, 12559–12571 (2020)

2020
[24]

Duvenaud, D. K. et al. Convolutional networks on graphs for learning molecular fingerprints. Advances in neural information processing systems 28 (2015)

2015
[25]

& Bjerrum, E

Irwin, R., Dimitriadis, S., He, J. & Bjerrum, E. J. Chemformer: a pre-trained transformer for computational chemistry. Machine Learning: Science and Technology 3, 015022 (2022). 16

2022
[26]

Agile platform: a deep learning powered approach to accelerate lnp development for mrna delivery

Xu, Y .et al. Agile platform: a deep learning powered approach to accelerate lnp development for mrna delivery. Nature communications 15, 6305 (2024)

2024
[27]

& Barati Farimani, A

Wang, Y ., Wang, J., Cao, Z. & Barati Farimani, A. Molecular contrastive learning of representations via graph neural networks. Nature Machine Intelligence 4, 279–287 (2022)

2022
[28]

& Hinton, G

Chen, T., Kornblith, S., Norouzi, M. & Hinton, G. A simple framework for contrastive learning of visual representations. In International conference on machine learning, 1597–1607 (PMLR, 2020)

2020
[29]

& Hahn, M

Rogers, D. & Hahn, M. Extended-connectivity fingerprints. Journal of Chemical Information and Modeling50, 742–754 (2010). URL https://doi.org/10.1021/ci100050t‘‘‘‘ . PMID: 20426451, https://doi.org/10.1021/ci100050t

2010 doi
[30]

Rdkit: Open-source cheminformatics

Landrum, G. Rdkit: Open-source cheminformatics. Online 1 (2013)

2013
[31]

& Landrum, G

Riniker, S. & Landrum, G. A. Better informed distance geometry: Using what we know to improve conformation generation. Journal of Chemical Information and Modeling 55, 2562–2574 (2015)

2015
[32]

& Hinton, G

Van der Maaten, L. & Hinton, G. Visualizing data using t-sne. Journal of machine learning research9 (2008)

2008
[33]

& Khanna, R

Awad, M., Khanna, R., Awad, M. & Khanna, R. Support vector regression. Efficient learning machines: Theories, concepts, and applications for engineers and system designers 67–80 (2015)

2015
[34]

& Schölkopf, B

Vert, J.-P., Tsuda, K. & Schölkopf, B. 2 a primer on kernel methods. Kernel Methods in Computational Biology 35 (2004)

2004
[35]

& Takagi, T

Moriwaki, H., Tian, Y .-S., Kawashita, N. & Takagi, T. Mordred: a molecular descriptor calculator.Journal of cheminformatics 10, 4 (2018)

2018
[36]

Vaswani, A., Shazeer, N. & et al. Attention is all you need. Advances in neural information processing systems 30 (2017)

2017
[37]

& Toutanova, K

Devlin, J., Chang, M.-W., Lee, K. & Toutanova, K. Bert: Pre-training of deep bidirectional transformers for language understanding (2019). 1810.04805

2019 arXiv
[38]

Random forests

Breiman, L. Random forests. Machine learning 45, 5–32 (2001)

2001
[39]

K-nearest neighbors

Kramer, O. K-nearest neighbors. Dimensionality reduction with unsupervised nearest neighbors 13–23 (2013)

2013
[40]

Heid, E. et al. Chemprop: A machine learning package for chemical property prediction. Journal of Chemical Information and Modeling 64, 9–17 (2024). URL https://doi.org/10.1021/acs.jcim.3c01250. PMID: 38147829, https://doi. org/10.1021/acs.jcim.3c01250

2024 doi
[41]

Li, H. et al. A knowledge-guided pre-training framework for improving molecular representation learning. Nature Communica- tions 14, 7568 (2023)

2023
[42]

& Lee, J

Belenahalli Shekarappa, S., Kandagalla, S. & Lee, J. Development of machine learning models based on molecular fingerprints for selection of small molecule inhibitors against jak2 protein. Journal of Computational Chemistry 44, 1493–1504 (2023)

2023
[43]

& Bajorath, J

Maggiora, G., V ogt, M., Stumpfe, D. & Bajorath, J. Molecular similarity in medicinal chemistry: miniperspective.Journal of medicinal chemistry 57, 3186–3204 (2014)

2014
[44]

Petrone, P. M. et al. Rethinking molecular similarity: comparing compounds on the basis of biological activity. ACS chemical biology 7, 1399–1409 (2012)

2012
[45]

& Rácz, A

Orosz, Á., Héberger, K. & Rácz, A. Comparison of descriptor-and fingerprint sets in machine learning models for adme-tox targets. Frontiers in Chemistry 10, 852893 (2022)

2022
[46]

Bioactivity assessment of natural compounds using machine learning models trained on target similarity between drugs

Periwal, V .et al. Bioactivity assessment of natural compounds using machine learning models trained on target similarity between drugs. PLoS computational biology 18, e1010029 (2022)

2022
[47]

& Tang, J

Zagidullin, B., Wang, Z., Guan, Y ., Pitkänen, E. & Tang, J. Comparative analysis of molecular fingerprints in prediction of drug combination effects. Briefings in bioinformatics 22, bbab291 (2021)

2021
[48]

Tejera, E. et al. Cell fishing: A similarity based approach and machine learning strategy for multiple cell lines-compound sensitivity prediction. PLoS One 14, e0223276 (2019)

2019
[49]

& Guan, X

Zhong, S. & Guan, X. Count-based morgan fingerprint: A more efficient and interpretable molecular representation in develop- ing machine learning-based predictive regression models for water contaminants’ activities and properties. Environmental science & technology 57, 18193–...

2023
[50]

Hoerl, A. E. & Kennard, R. W. Ridge regression: Biased estimation for nonorthogonal problems. Technometrics 12, 55–67 (1970)

1970
[51]

& Scholkopf, B

Hearst, M., Dumais, S., Osuna, E., Platt, J. & Scholkopf, B. Support vector machines. IEEE Intelligent Systems and their Applications 13, 18–28 (1998)

1998
[52]

& Guestrin, C

Chen, T. & Guestrin, C. Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, 785–794 (2016)

2016
[53]

Prokhorenkova, L., Gusev, G., V orobev, A., Dorogush, A. V . & Gulin, A. Catboost: unbiased boosting with categorical features. Advances in neural information processing systems 31 (2018)

2018
[54]

& Skolnick, J

Zhou, H. & Skolnick, J. Utility of the morgan fingerprint in structure-based virtual ligand screening. The Journal of Physical Chemistry B 128, 5363–5370 (2024). 17

2024
[55]

Tayyebi, A. et al. Prediction of organic compound aqueous solubility using machine learning: a comparison study of descriptor-based and fingerprints-based models. Journal of Cheminformatics 15, 99 (2023)

2023
[56]

& Takagi, T

Moriwaki, H., Tian, Y .-S., Kawashita, N. & Takagi, T. Moesm4 of mordred: a molecular descriptor calculator. (No Title) (2018)

2018
[57]

& Bender, A

Seal, S., Yang, H., V ollmers, L. & Bender, A. Comparison of cellular morphological descriptors and molecular fingerprints for the prediction of cytotoxicity-and proliferation-related assays. Chemical Research in Toxicology34, 422–437 (2021)

2021
[58]

& Liu, S

Fang, S., Liu, Y . & Liu, S. Mfgb: molecular properties prediction leveraging self-supervised morgan fingerprint representation learning. In 2023 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 3004–3011 (IEEE, 2023)

2023
[59]

Gaulton, A. et al. The chembl database in 2017. Nucleic acids research 45, D945–D954 (2017)

2017
[60]

Mehradfar, A. et al. Supervised learning for analog and rf circuit design: Benchmarks and comparative insights. arXiv preprint arXiv:2501.11839 (2025)

2025 arXiv
[61]

& Kang, J

An, Q., Rahman, S., Zhou, J. & Kang, J. J. A comprehensive review on machine learning in healthcare industry: classification, restrictions, opportunities and challenges. Sensors 23, 4178 (2023)

2023
[62]

Kingma, D. P. & Ba, J. Adam: A method for stochastic optimization (2017). 1412.6980

2017 arXiv
[63]

Ramsundar, B. et al. Deep Learning for the Life Sciences (O’Reilly Media, 2019). https://www.amazon.com/ Deep-Learning-Life-Sciences-Microscopy/dp/1492039837 . 18 Supplementary Information (a) AGILE with original data (full set) (b) AGILE with original data (test split) (c) AG...

2019

Pith tools

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

[1] [1]

Non-viral vectors for rna delivery

Yan, Y .et al. Non-viral vectors for rna delivery. Journal of Controlled Release 342, 241–279 (2022)

work page 2022

[2] [2]

& Jin, H

Zhu, Y ., Zhu, L., Wang, X. & Jin, H. Rna-based therapeutics: an overview and prospectus.Cell death & disease 13, 644 (2022)

work page 2022

[3] [3]

L., Ciccone, G., Esposito, C

Ibba, M. L., Ciccone, G., Esposito, C. L., Catuogno, S. & Giangrande, P. H. Advances in mrna non-viral delivery approaches. Advanced drug delivery reviews 177, 113930 (2021)

work page 2021

[4] [4]

Kim, M. et al. Engineered ionizable lipid nanoparticles for targeted delivery of rna therapeutics into different types of cells in the liver. Science advances 7, eabf4398 (2021)

work page 2021

[5] [5]

Wittrup, A. et al. Visualizing lipid-formulated sirna release from endosomes and target gene knockdown. Nature biotechnology 33, 870–876 (2015)

work page 2015

[6] [6]

Xu, E., Saltzman, W. M. & Piotrowski-Daspit, A. S. Escaping the endosome: assessing cellular trafficking mechanisms of non-viral vehicles. Journal of Controlled Release 335, 465–480 (2021)

work page 2021

[7] [7]

Albertsen, C. H. et al. The role of lipid components in lipid nanoparticles for vaccines and gene therapy. Advanced drug delivery reviews 188, 114416 (2022)

work page 2022

[8] [8]

Cheng, Q. et al. Selective organ targeting (sort) nanoparticles for tissue-specific mrna delivery and crispr–cas gene editing. Nature nanotechnology 15, 313–320 (2020)

work page 2020