Pith. sign in

REVIEW 3 major objections 5 minor 85 references

Programmable Virtual Humans Toward Human Physiologically-Based Drug Discovery

T0 review · 3 major / 5 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper argues that recent advances make it possible to construct programmable virtual humans that simulate drug action across the whole human body.

desk verdict A clear, honest vision paper that names the OOD bottleneck, but the feasibility claim is asserted not demonstrated and the self-citation pattern detracts. read the letter →

arxiv 2507.19568 v1 pith:GTU27HDP submitted 2025-07-25 cs.CY cs.AIcs.CEcs.LG

You Wu , Philip E. Bourne , Lei Xie This is my paper
classification cs.CYcs.AIcs.CEcs.LG
keywords programmablevirtualhumansphysiologicallybaseddrugdiscoverydigitaltwinsout-of-distributionpredictionmulti-omicsintegrationsingle-cellomicsuncertaintyquantificationpharmacokinetics
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 perspective proposes a new paradigm for drug discovery built around a programmable virtual human (PVH): a dynamic, multiscale computer model that simulates what a novel molecule does in the human body from molecular interactions up to clinical phenotypes. The authors argue that current AI tools only digitize existing experiments, while medical digital twins lack the resolution for early-stage discovery, leaving a translational gap that drives clinical failures. They contend that recent progress in AI, high-throughput perturbation assays, and single-cell and spatial omics now makes building PVHs feasible. If realized, PVHs would let researchers screen and optimize drug candidates directly for human efficacy and safety in silico, before any human trial.

What carries the argument

The central object is the Programmable Virtual Human (PVH), a dynamic, multiscale model that simulates drug actions from molecular to phenotypic levels. Its machinery combines data-driven machine learning with mechanism-based approaches: physics-informed neural networks and PBPK/QSP models for pharmacokinetics, genome-wide protein-compound interaction predictors for target engagement, single-cell and spatial omics models for cell state responses, and multi-scale integration to link molecular perturbations to clinical phenotypes. The authors propose that causal representation learning, uncertainty quantification, and integration of mechanism-based models will make PVHs generalizable to out-of-distribution molecules and patients.

What would settle it

A prospective benchmark in which a PVH-like model, trained only on pre-clinical data from cell lines, organoids, animal models, and biobanks, must predict the clinical efficacy and safety of ten or more new chemical entities before any human trial data exists; if its rankings of candidates correlate no better than chance with observed clinical outcomes, the central feasibility claim would be refuted.

Watch

Extended reading notes

Core claim

The paper's central claim is that the convergence of deep learning, chemical perturbation screens, single-cell and spatial omics, and mechanism-based modeling makes it possible to build programmable virtual humans: integrated computational models that simulate the full life cycle of a drug in the body, from absorption and distribution to molecular target engagement, cellular phenotype changes, and organism-level clinical outcomes. Unlike existing digital twins that depend on human clinical data, a PVH would predict the effects of new, unseen molecules using only pre-clinical and model-organism data, enabling direct optimization of therapeutic efficacy and safety and inverse design of molecules that revert a disease state to a healthy state.

Load-bearing premise

The premise that molecular and cellular data from cell lines, organoids, animals, and biobanks can be combined to reliably predict how a brand-new molecule will act in an actual human body.

Editorial extensions

If this is right

  • Drug candidates could be screened and optimized by simulating their clinical outcomes in a virtual human rather than by affinity or potency in isolated assays.
  • The approach could close the translational gap between early discovery and late-phase development, potentially reducing the high failure rate of drugs in clinical trials.
  • Inverse design becomes possible: given a patient's disease state and a healthy reference state as prompts, a generative model could propose molecules that revert the disease phenotype to health.
  • Model systems such as cell lines, organoids, and animal studies would be used to train and calibrate PVHs rather than as direct proxies for human response.
  • PVHs could unify data and process silos across drug discovery, linking molecular-level models with physiological and real-world clinical models.

Reading between the lines

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

  • If PVHs prove reliable for well-characterized molecules, a natural next step would be prospective benchmarking against historical clinical trial failures to measure how many late-stage failures could have been predicted in silico.
  • The same multiscale framework could extend beyond small-molecule drugs to biologics, gene therapies, and combination regimens, since the central challenge is predicting systemic human response to a novel perturbation.
  • A testable intermediate goal is whether a PVH can predict organ-level adverse effects, such as cardiotoxicity, from single-cell perturbation data alone, without human trial data for the compound.
  • The heavy reliance on out-of-distribution generalization suggests that progress on PVHs may be gated more by fundamental advances in causal representation learning and uncertainty quantification than by additional omics data alone.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. The manuscript is a perspective article proposing 'Programmable Virtual Humans' (PVH): dynamic, multiscale computational models intended to simulate the complete pharmaceutical action of novel molecules in the human body, from molecular interactions through cellular and tissue responses to clinical phenotypes. The authors argue that current AI-based drug discovery digitizes isolated experimental steps and thereby fails to bridge the translational gap between model systems and humans, whereas PVH would enable direct in silico optimization of therapeutic efficacy and safety. The paper surveys relevant data sources (single-cell, spatial, and chemoproteomic omics; perturbation screens; MPS), discusses modeling approaches (deep learning, PINNs, causal representation learning, mechanism-based models), and proposes a three-pronged roadmap addressing out-of-distribution (OOD) generalization, uncertainty quantification, and multi-scale integration. The central claim is that recent advances make PVH construction possible, but the paper explicitly acknowledges that the OOD space is 'staggeringly vast' and that existing methods fail in small-data, zero-shot settings, leaving feasibility as an open research agenda rather than a demonstrated capability.

Significance. If realized, PVH would indeed represent a paradigm shift in drug discovery, moving optimization from target binding or cell-line potency to predicted human physiological outcomes. The paper's main contribution is conceptual: it clearly articulates the limitations of current AI-for-drug-discovery pipelines, identifies the specific translational gaps, and organizes a plausible research program around OOD learning, uncertainty quantification, and mechanism-based integration. A notable strength is its honesty about the obstacles: Section 4.1 concedes the vastness of the OOD space and the failure of conventional machine learning in zero-shot and what-if settings, which sets the paper above uncritical advocacy. However, the paper provides no demonstration, benchmark, or quantitative feasibility analysis supporting the premise that the proposed approaches can close the distribution-shift gap; the cited advances are ingredient technologies, not an integrated working system. As a perspective, the paper is useful and internally coherent, but the central feasibility claim is asserted rather than established.

major comments (3)
  1. [Abstract; §4.1–4.3] The claim that recent advances 'now make it possible' to construct PVH is the load-bearing assertion, but §4.1 states that the OOD space is 'staggeringly vast' and that conventional ML fails in small-data, zero-shot, and what-if settings, and §4.2–4.3 present causal representation learning, new uncertainty quantification, and mechanism-based integration only as research directions with no worked example, benchmark, or quantitative argument that they close the distribution-shift gap. To make the central claim defensible, the paper should either temper the abstract and conclusion to describe a research agenda 'toward' PVH, or specify a minimal-viable PVH with concrete, falsifiable success criteria.
  2. [§4.3] The argument that mechanism-based modeling improves generalizability is weakened by the paper's own enumeration of its limitations: incomplete mechanism knowledge, parameterization uncertainty, computational intractability, and reliance on simplifying assumptions. The proposed integration with machine learning is plausible, but the cited examples (metabolic-model-informed classifiers, PINNs in climate modeling) do not demonstrate that the hybrid overcomes these limitations for novel molecules in a whole-person context. Please add at least one concrete proof-of-concept or benchmark showing that a hybrid mechanistic/data-driven model generalizes to unseen chemical or patient distributions.
  3. [§3.2–3.3, §4.4] The roadmap relies on translating cross-species and in vitro data to human responses through multi-omics profiling and cross-level end-to-end learning, but the cited successes are predominantly in silico or cell-line benchmarks, and the paper offers no evidence of prospective clinical validation. For a perspective this is acceptable only if the text explicitly frames the translational leap as an open question rather than a near-term opportunity; currently phrases such as 'it becomes possible to reverse-engineer the complexities of human physiology' overstate the demonstrated capability. The paper should state what evidence would count as translational success, for example prospective concordance between PVH predictions and clinical trial outcomes.
minor comments (5)
  1. [Author footnote] The word 'Correponding' should be 'Corresponding'.
  2. [Section 2] The text contains 'GW AS' with an erroneous space; it should be 'GWAS'.
  3. [Section 4.1] The heading 'Challenges in developing the PHV' uses 'PHV' while the rest of the paper uses 'PVH'; the same inconsistency appears in §4.2 ('multi-modal embedding space of the PHV').
  4. [Conclusion] The word 'intepretability' should be 'interpretability', and in Section 4.4 'embeded' should be 'embedded'.
  5. [Introduction] 'pay little attentions' should be 'pay little attention'.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper is a forward-looking perspective whose feasibility claim is asserted and supported by cited ingredients, not derived from them.

full rationale

This paper is a perspective/roadmap, not a derivation. It contains no equations fitted to data, no parameter estimation, and no numerical prediction that could be equivalent to an input by construction. The central concept, programmable virtual humans (PVH), is introduced as a new proposal rather than being defined via any output it is said to predict; likewise, the proposal's expected benefits (Section 2) are not used as inputs to any model. The paper's own Section 4.1 candidly concedes that the OOD space is 'staggeringly vast' and that conventional ML 'cannot handle small data or OOD cases' and fails in zero-shot and 'what-if' settings, and Sections 4.2-4.3 present causal representation learning, new uncertainty quantification, and mechanism-based integration as research directions rather than as accomplished reductions of the gap. The many self-citations (e.g., refs 10, 19, 47, 61, 65, 67, 79) are used as examples of ingredient technologies that 'paved the way' and 'hold significant potential'; they do not by themselves entail that PVH is achievable, and the paper does not invoke any uniqueness theorem or forbid alternatives by citing the authors' own prior work. Removing those citations would leave the perspective's thesis intact as a proposal. Hence no load-bearing step reduces to its own input, and no self-citation chain forces the conclusion. This is the common honest non-finding: the feasibility claim is under-supported but not circular.

Assumptions & free parameters 0 free parameters · 4 assumptions · 1 invented entities

The paper rests on feasibility assumptions about data availability, OOD generalization, and multi-scale integration. No free parameters are fitted because no quantitative model is presented. The PVH itself is an invented conceptual entity without independent evidence.

assumptions (4)
  • domain assumption Recent advances in AI, high-throughput perturbation assays, and single-cell and spatial omics across species now make it possible to construct programmable virtual humans.
    Stated in the Abstract and expanded in Section 3; it is the feasibility premise of the whole paper, relying on the availability and integrability of these data types at scale.
  • domain assumption With sufficient perturbation data, it becomes possible to reverse-engineer the complexities of human physiology.
    Section 3.3, paragraph 1. Assumes that perturbation screens (perturb-seq, drug-seq, cell painting, MPS) provide enough signal to reconstruct physiological response mechanisms.
  • domain assumption Causal representation learning enables models to generalize better to new, unseen distributions by learning invariant causal factors.
    Section 4.2. Assumes OOD generalization in drug discovery is achievable via causal representation learning, a claim not demonstrated for the scale of chemistry-proteome-phenotype space.
  • domain assumption Mechanism-based models can make predictions by leveraging existing knowledge, even in situations where data is scarce.
    Section 4.3. Assumes mechanistic models (e.g., constraint-based metabolic models, PBPK) can compensate for small data and distribution shift at the whole-person scale.
invented entities (1)
  • Programmable Virtual Human (PVH)
    purpose: A proposed dynamic, multiscale computational model simulating drug action from molecular interactions to clinical phenotypes, intended for early-stage drug discovery and inverse design.
    Introduced in Sections 1 and 2; no implementation, specification, or falsifiable prediction is provided, so there is no evidence outside the paper that such a system can be built beyond the cited partial components.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Programmable Virtual Humans Toward Human Physiologically-Based Drug Discovery." pith.science (2026). https://pith.science/paper/GTU27HDP

@misc{pith2026250719568,
  author       = {Pith},
  title        = {Pith review of: Programmable Virtual Humans Toward Human Physiologically-Based Drug Discovery},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GTU27HDP}},
  note         = {Machine review of arXiv:2507.19568}
}
read the original abstract

Artificial intelligence (AI) has sparked immense interest in drug discovery, but most current approaches only digitize existing high-throughput experiments. They remain constrained by conventional pipelines. As a result, they do not address the fundamental challenges of predicting drug effects in humans. Similarly, biomedical digital twins, largely grounded in real-world data and mechanistic models, are tailored for late-phase drug development and lack the resolution to model molecular interactions or their systemic consequences, limiting their impact in early-stage discovery. This disconnect between early discovery and late development is one of the main drivers of high failure rates in drug discovery. The true promise of AI lies not in augmenting current experiments but in enabling virtual experiments that are impossible in the real world: testing novel compounds directly in silico in the human body. Recent advances in AI, high-throughput perturbation assays, and single-cell and spatial omics across species now make it possible to construct programmable virtual humans: dynamic, multiscale models that simulate drug actions from molecular to phenotypic levels. By bridging the translational gap, programmable virtual humans offer a transformative path to optimize therapeutic efficacy and safety earlier than ever before. This perspective introduces the concept of programmable virtual humans, explores their roles in a new paradigm of drug discovery centered on human physiology, and outlines key opportunities, challenges, and roadmaps for their realization.

Figures

Figures reproduced from arXiv: 2507.19568 by the authors.

Figure 1
Figure 1. (A) Overview of three drug discovery paradigms: target-based, phenotype￾based, and physiology-based. Each leverages distinct information, such as protein activity, imaging, or multi-omics, to build computational models (e.g., QSAR, image processing, digital twins) with objectives ranging from binding affinity prediction to morphological changes and clinical safety assessment. (B) The proposed ”PVH (Pro￾grammable Vir… view at source ↗

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

85 extracted references · 76 canonical work pages

  1. [1]

    Target-based drug discovery: is something wrong?,

    F. Sams-Dodd, “Target-based drug discovery: is something wrong?,” Drug discovery today, vol. 10, no. 2, pp. 139–147, 2005

    2005

  2. [2]

    Artificial intelligence in drug discovery: what is realistic, what are illusions? part 1: Ways to make an impact, and why we are not there yet,

    A. Bender and I. Cort´ es-Ciriano, “Artificial intelligence in drug discovery: what is realistic, what are illusions? part 1: Ways to make an impact, and why we are not there yet,”Drug discovery today, vol. 26, no. 2, pp. 511–524, 2021

    2021

  3. [3]

    Digital twins to personalize medicine,

    B. Bj¨ ornsson, C. Borrebaeck, N. Elander, T. Gasslander, D. R. Gawel, M. Gustafsson, R. J¨ ornsten, E. J. Lee, X. Li, S. Lilja,et al., “Digital twins to personalize medicine,” Genome medicine, vol. 12, pp. 1–4, 2020

    2020

  4. [4]

    Digital twins in medicine,

    R. Laubenbacher, B. Mehrad, I. Shmulevich, and N. Trayanova, “Digital twins in medicine,” Nature Computational Science , vol. 4, no. 3, pp. 184– 191, 2024

    2024

  5. [5]

    Digital twins in healthcare: is it the beginning of a new era of evidence- based medicine? a critical review,

    P. Armeni, I. Polat, L. M. De Rossi, L. Diaferia, S. Meregalli, and A. Gatti, “Digital twins in healthcare: is it the beginning of a new era of evidence- based medicine? a critical review,” Journal of Personalized Medicine , vol. 12, no. 8, p. 1255, 2022

    2022

  6. [6]

    A roadmap for the development of human body digital twins,

    C. Tang, W. Yi, E. Occhipinti, Y. Dai, S. Gao, and L. G. Occhipinti, “A roadmap for the development of human body digital twins,” Nature Reviews Electrical Engineering, vol. 1, no. 3, pp. 199–207, 2024

    work page 2024

  7. [7]

    A proteome-wide atlas of drug mechanism of action,

    D. C. Mitchell, M. Kuljanin, J. Li, J. G. Van Vranken, N. Bulloch, D. K. Schweppe, E. L. Huttlin, and S. P. Gygi, “A proteome-wide atlas of drug mechanism of action,” Nature Biotechnology, vol. 41, no. 6, pp. 845–857, 2023

    work page 2023

  8. [8]

    Chemical proteomics reveals the target landscape of 1,000 kinase inhibitors,

    M. Reinecke, P. Brear, L. Vornholz, B.-T. Berger, F. Seefried, S. Wilhelm, P. Samaras, L. Gyenis, D. W. Litchfield, G. M´ edard, et al. , “Chemical proteomics reveals the target landscape of 1,000 kinase inhibitors,” Nature Chemical Biology, vol. 20, no. 5, pp. 577–585, 2024

    work page 2024

Show all 85 references
  1. [9]

    Large-scale chemoproteomics expedites ligand discovery and predicts lig- and behavior in cells,

    F. Offensperger, G. Tin, M. Duran-Frigola, E. Hahn, S. Dobner, C. W. a. Ende, J. W. Strohbach, A. Rukavina, V. Brennsteiner, K. Ogilvie, et al. , “Large-scale chemoproteomics expedites ligand discovery and predicts lig- and behavior in cells,” Science, vol. 384, no. 6694, p. e...

    2024

  2. [10]

    Ai-driven multi-omics integration for multi-scale pre- dictive modeling of genotype-environment-phenotype relationships,

    Y. Wu and L. Xie, “Ai-driven multi-omics integration for multi-scale pre- dictive modeling of genotype-environment-phenotype relationships,” Com- putational and Structural Biotechnology Journal , 2025

    2025

  3. [11]

    Applications, challenges, and outlook for pbpk modeling and simulation: a regulatory, industrial and academic perspective,

    W. Lin, Y. Chen, J. D. Unadkat, X. Zhang, D. Wu, and T. Heimbach, “Applications, challenges, and outlook for pbpk modeling and simulation: a regulatory, industrial and academic perspective,” Pharmaceutical Research, vol. 39, no. 8, pp. 1701–1731, 2022. 13

    2022

  4. [12]

    Machine learning and artificial intelligence in physiologically based pharmacokinetic modeling,

    W.-C. Chou and Z. Lin, “Machine learning and artificial intelligence in physiologically based pharmacokinetic modeling,” Toxicological Sciences, vol. 191, no. 1, pp. 1–14, 2023

    2023

  5. [13]

    Mole: a foundation model for molecular graphs using disentangled attention,

    O. M´ endez-Lucio, C. A. Nicolaou, and B. Earnshaw, “Mole: a foundation model for molecular graphs using disentangled attention,” Nature Commu- nications, vol. 15, no. 1, p. 9431, 2024

    2024

  6. [14]

    Scientific machine learning through physics–informed neural net- works: Where we are and what’s next,

    S. Cuomo, V. S. Di Cola, F. Giampaolo, G. Rozza, M. Raissi, and F. Pic- cialli, “Scientific machine learning through physics–informed neural net- works: Where we are and what’s next,” Journal of Scientific Computing , vol. 92, no. 3, p. 88, 2022

    2022

  7. [15]

    Climode: Climate and weather forecasting with physics-informed neural odes,

    Y. Verma, M. Heinonen, and V. Garg, “Climode: Climate and weather forecasting with physics-informed neural odes,” arXiv preprint arXiv:2404.10024, 2024

    2024 arXiv

  8. [16]

    Novel computational approaches to polypharmacology as a means to define responses to individ- ual drugs,

    L. Xie, L. Xie, S. L. Kinnings, and P. E. Bourne, “Novel computational approaches to polypharmacology as a means to define responses to individ- ual drugs,” Annual review of pharmacology and toxicology , vol. 52, no. 1, pp. 361–379, 2012

    2012

  9. [17]

    Evolutionary-scale prediction of atomic- level protein structure with a language model,

    Z. Lin, H. Akin, R. Rao, B. Hie, Z. Zhu, W. Lu, N. Smetanin, R. Verkuil, O. Kabeli, Y. Shmueli, et al. , “Evolutionary-scale prediction of atomic- level protein structure with a language model,” Science, vol. 379, no. 6637, pp. 1123–1130, 2023

    2023

  10. [18]

    Accurate struc- ture prediction of biomolecular interactions with alphafold 3,

    J. Abramson, J. Adler, J. Dunger, R. Evans, T. Green, A. Pritzel, O. Ron- neberger, L. Willmore, A. J. Ballard, J. Bambrick, et al., “Accurate struc- ture prediction of biomolecular interactions with alphafold 3,” Nature, pp. 1–3, 2024

    2024

  11. [19]

    Model agnostic semi-supervised meta-learning elucidates understudied out-of-distribution molecular inter- actions,

    Y. Wu, L. Xie, Y. Liu, and L. Xie, “Model agnostic semi-supervised meta-learning elucidates understudied out-of-distribution molecular inter- actions,” bioRxiv, 2024

    2024

  12. [20]

    End-to-end sequence- structure-function meta-learning predicts genome-wide chemical-protein in- teractions for dark proteins,

    T. Cai, L. Xie, S. Zhang, M. Chen, D. He, A. Badkul, Y. Liu, H. K. Namballa, M. Dorogan, W. W. Harding, et al. , “End-to-end sequence- structure-function meta-learning predicts genome-wide chemical-protein in- teractions for dark proteins,” PLOS Computational Biology , vol. 19...

    2023

  13. [21]

    Leveraging transfer learning for predicting protein-small molecule interactions,

    J. Wang and N. V. Dokholyan, “Leveraging transfer learning for predicting protein-small molecule interactions,” bioRxiv, pp. 2024–10, 2024

    2024

  14. [22]

    Contrastive learning in protein language space predicts interactions between drugs and protein targets,

    R. Singh, S. Sledzieski, B. Bryson, L. Cowen, and B. Berger, “Contrastive learning in protein language space predicts interactions between drugs and protein targets,” Proceedings of the National Academy of Sciences, vol. 120, no. 24, p. e2220778120, 2023. 14

    2023

  15. [23]

    Toward high-throughput predictive modeling of protein binding/unbinding kinetics,

    S. H. Chiu and L. Xie, “Toward high-throughput predictive modeling of protein binding/unbinding kinetics,” Journal of chemical information and modeling, vol. 56, no. 6, pp. 1164–1174, 2016

    2016

  16. [24]

    A euclidean transformer for fast and stable machine learned force fields,

    J. T. Frank, O. T. Unke, K.-R. M¨ uller, and S. Chmiela, “A euclidean transformer for fast and stable machine learned force fields,” Nature Com- munications, vol. 15, no. 1, p. 6539, 2024

    2024

  17. [25]

    Direct generation of protein conformational ensembles via machine learning,

    G. Janson, G. Valdes-Garcia, L. Heo, and M. Feig, “Direct generation of protein conformational ensembles via machine learning,” Nature Commu- nications, vol. 14, no. 1, p. 774, 2023

    2023

  18. [26]

    Deepreal: a deep learning powered multi-scale modeling framework for predicting out-of-distribution ligand- induced gpcr activity,

    T. Cai, K. A. Abbu, Y. Liu, and L. Xie, “Deepreal: a deep learning powered multi-scale modeling framework for predicting out-of-distribution ligand- induced gpcr activity,” Bioinformatics, vol. 38, no. 9, pp. 2561–2570, 2022

    2022

  19. [27]

    Single-cell genomics meets human genetics,

    A. S. Cuomo, A. Nathan, S. Raychaudhuri, D. G. MacArthur, and J. E. Powell, “Single-cell genomics meets human genetics,” Nature Reviews Ge- netics, vol. 24, no. 8, pp. 535–549, 2023

    2023

  20. [28]

    Single-cell epigenomics: Recording the past and predicting the future,

    G. Kelsey, O. Stegle, and W. Reik, “Single-cell epigenomics: Recording the past and predicting the future,” Science, vol. 358, no. 6359, pp. 69–75, 2017

    2017

  21. [29]

    Single-cell rna sequencing technologies and applications: A brief overview,

    D. Jovic, X. Liang, H. Zeng, L. Lin, F. Xu, and Y. Luo, “Single-cell rna sequencing technologies and applications: A brief overview,” Clinical and translational medicine, vol. 12, no. 3, p. e694, 2022

    2022

  22. [30]

    Single-cell proteomics enabled by next-generation sequencing or mass spectrometry,

    H. M. Bennett, W. Stephenson, C. M. Rose, and S. Darmanis, “Single-cell proteomics enabled by next-generation sequencing or mass spectrometry,” Nature Methods, vol. 20, no. 3, pp. 363–374, 2023

    2023

  23. [31]

    Insights into protein post- translational modification landscapes of individual human cells by trapped ion mobility time-of-flight mass spectrometry,

    B. C. Orsburn, Y. Yuan, and N. N. Bumpus, “Insights into protein post- translational modification landscapes of individual human cells by trapped ion mobility time-of-flight mass spectrometry,” Nature communications , vol. 13, no. 1, p. 7246, 2022

    2022

  24. [32]

    Spatial multimodal analysis of transcriptomes and metabolomes in tissues,

    M. Vicari, R. Mirzazadeh, A. Nilsson, R. Shariatgorji, P. Bj¨ arterot, L. Lars- son, H. Lee, M. Nilsson, J. Foyer, M. Ekvall, et al. , “Spatial multimodal analysis of transcriptomes and metabolomes in tissues,” Nature Biotech- nology, vol. 42, no. 7, pp. 1046–1050, 2024

    2024

  25. [33]

    Single-cell lipidomics with high structural specificity by mass spectrometry,

    Z. Li, S. Cheng, Q. Lin, W. Cao, J. Yang, M. Zhang, A. Shen, W. Zhang, Y. Xia, X. Ma, et al., “Single-cell lipidomics with high structural specificity by mass spectrometry,” Nature communications, vol. 12, no. 1, p. 2869, 2021

    2021

  26. [34]

    Native n-glycome profiling of single cells and ng-level blood isolates using label-free capillary electrophoresis- mass spectrometry,

    A.-L. Marie, Y. Gao, and A. R. Ivanov, “Native n-glycome profiling of single cells and ng-level blood isolates using label-free capillary electrophoresis- mass spectrometry,” Nature Communications, vol. 15, no. 1, p. 3847, 2024. 15

    2024

  27. [35]

    Single-cell transcriptomics across 2,534 microbial species reveals functional heterogeneity in the rumen microbiome,

    M. Jia, S. Zhu, M.-Y. Xue, H. Chen, J. Xu, M. Song, Y. Tang, X. Liu, Y. Tao, T. Zhang, et al., “Single-cell transcriptomics across 2,534 microbial species reveals functional heterogeneity in the rumen microbiome,” Nature Microbiology, pp. 1–15, 2024

    2024

  28. [36]

    Single-cell morphological and topolog- ical atlas reveals the ecosystem diversity of human breast cancer,

    S. Zhao, D.-P. Chen, T. Fu, J.-C. Yang, D. Ma, X.-Z. Zhu, X.-X. Wang, Y.-P. Jiao, X. Jin, Y. Xiao, et al. , “Single-cell morphological and topolog- ical atlas reveals the ecosystem diversity of human breast cancer,” Nature Communications, vol. 14, no. 1, p. 6796, 2023

    2023

  29. [37]

    Supervised discov- ery of interpretable gene programs from single-cell data,

    R. Z. Kunes, T. Walle, M. Land, T. Nawy, and D. Pe’er, “Supervised discov- ery of interpretable gene programs from single-cell data,” Nature biotech- nology, vol. 42, no. 7, pp. 1084–1095, 2024

    2024

  30. [38]

    Single-cell and spatial multiomic inference of gene regulatory networks using scripro,

    Z. Chang, Y. Xu, X. Dong, Y. Gao, and C. Wang, “Single-cell and spatial multiomic inference of gene regulatory networks using scripro,” Bioinfor- matics, vol. 40, no. 7, p. btae466, 2024

    2024

  31. [39]

    D’or: Deep orienter of protein-protein interaction networks,

    D. Pirak and R. Sharan, “D’or: Deep orienter of protein-protein interaction networks,” Bioinformatics, p. btae355, 2024

    2024

  32. [40]

    Machine learn- ing for the advancement of genome-scale metabolic modeling,

    P. Kundu, S. Beura, S. Mondal, A. K. Das, and A. Ghosh, “Machine learn- ing for the advancement of genome-scale metabolic modeling,” Biotechnol- ogy Advances, p. 108400, 2024

    2024

  33. [41]

    Biophysical mod- eling with variational autoencoders for bimodal, single-cell rna sequencing data,

    M. Carilli, G. Gorin, Y. Choi, T. Chari, and L. Pachter, “Biophysical mod- eling with variational autoencoders for bimodal, single-cell rna sequencing data,” Nature Methods, vol. 21, no. 8, pp. 1466–1469, 2024

    2024

  34. [42]

    Multimodal data fusion for cancer biomarker discovery with deep learning,

    S. Steyaert, M. Pizurica, D. Nagaraj, P. Khandelwal, T. Hernandez- Boussard, A. J. Gentles, and O. Gevaert, “Multimodal data fusion for cancer biomarker discovery with deep learning,” Nature machine intelli- gence, vol. 5, no. 4, pp. 351–362, 2023

    2023

  35. [43]

    Sequence modeling and design from molecular to genome scale with evo,

    E. Nguyen, M. Poli, M. G. Durrant, B. Kang, D. Katrekar, D. B. Li, L. J. Bartie, A. W. Thomas, S. H. King, G. Brixi, et al. , “Sequence modeling and design from molecular to genome scale with evo,” Science, vol. 386, no. 6723, p. eado9336, 2024

    2024

  36. [44]

    The de- velopment of ai foundation models for single-cell transcriptomics,

    H. Bian, Y. Chen, E. Luo, X. Wu, M. Hao, L. Wei, and X. Zhang, “The de- velopment of ai foundation models for single-cell transcriptomics,”National Science Review, p. nwae340, 2024

    2024

  37. [45]

    Toward universal cell embeddings: integrating single-cell rna-seq datasets across species with saturn,

    Y. Rosen, M. Brbi´ c, Y. Roohani, K. Swanson, Z. Li, and J. Leskovec, “Toward universal cell embeddings: integrating single-cell rna-seq datasets across species with saturn,” Nature Methods, pp. 1–9, 2024

    2024

  38. [46]

    A cross-level information transmission network for hi- erarchical omics data integration and phenotype prediction from a new genotype,

    D. He and L. Xie, “A cross-level information transmission network for hi- erarchical omics data integration and phenotype prediction from a new genotype,” Bioinformatics, vol. 38, no. 1, pp. 204–210, 2022. 16

    2022

  39. [47]

    Hierarchical multi-omics data integration and modeling predict cell-specific chemical proteomics and drug responses,

    Y. Wu, Q. Liu, and L. Xie, “Hierarchical multi-omics data integration and modeling predict cell-specific chemical proteomics and drug responses,” Cell Reports Methods , vol. 3, no. 4, 2023

    2023

  40. [48]

    The central nervous system and the gut microbiome,

    G. Sharon, T. R. Sampson, D. H. Geschwind, and S. K. Mazmanian, “The central nervous system and the gut microbiome,” Cell, vol. 167, no. 4, pp. 915–932, 2016

    2016

  41. [49]

    The link between chronic kidney disease and cardiovascular disease,

    S. Said and G. T. Hernandez, “The link between chronic kidney disease and cardiovascular disease,” Journal of nephropathology, vol. 3, no. 3, p. 99, 2014

    2014

  42. [50]

    The dawn of spatial omics,

    D. Bressan, G. Battistoni, and G. J. Hannon, “The dawn of spatial omics,” Science, vol. 381, no. 6657, p. eabq4964, 2023

    2023

  43. [51]

    Multi-organ single-cell analy- sis reveals an on/off switch system with potential for personalized treatment of immunological diseases,

    S. Lilja, X. Li, M. Smelik, E. J. Lee, J. Loscalzo, P. B. Marthanda, L. Hu, M. Magnusson, O. Sysoev, H. Zhang, et al., “Multi-organ single-cell analy- sis reveals an on/off switch system with potential for personalized treatment of immunological diseases,” Cell Reports Medicin...

    2023

  44. [52]

    Programmable biology through artificial intelligence: from nucleic acids to proteins to cells,

    O. O. Abudayyeh and J. S. Gootenberg, “Programmable biology through artificial intelligence: from nucleic acids to proteins to cells,” Nature Meth- ods, vol. 21, no. 8, pp. 1384–1386, 2024

    2024

  45. [53]

    Dissecting key regulators of transcriptome kinetics through scalable single-cell rna profiling of pooled crispr screens,

    Z. Xu, A. Sziraki, J. Lee, W. Zhou, and J. Cao, “Dissecting key regulators of transcriptome kinetics through scalable single-cell rna profiling of pooled crispr screens,” Nature Biotechnology, vol. 42, no. 8, pp. 1218–1223, 2024

    2024

  46. [54]

    Scalable ge- netic screening for regulatory circuits using compressed perturb-seq,

    D. Yao, L. Binan, J. Bezney, B. Simonton, J. Freedman, C. J. Frangieh, K. Dey, K. Geiger-Schuller, B. Eraslan, A. Gusev, et al. , “Scalable ge- netic screening for regulatory circuits using compressed perturb-seq,” Na- ture biotechnology, vol. 42, no. 8, pp. 1282–1295, 2024

    2024

  47. [55]

    Epigenome editing technologies for discovery and medicine,

    S. R. McCutcheon, D. Rohm, N. Iglesias, and C. A. Gersbach, “Epigenome editing technologies for discovery and medicine,” Nature Biotechnology , vol. 42, no. 8, pp. 1199–1217, 2024

    2024

  48. [56]

    Drug-seq for miniaturized high- throughput transcriptome profiling in drug discovery,

    C. Ye, D. J. Ho, M. Neri, C. Yang, T. Kulkarni, R. Randhawa, M. Henault, N. Mostacci, P. Farmer, S. Renner, et al., “Drug-seq for miniaturized high- throughput transcriptome profiling in drug discovery,” Nature communica- tions, vol. 9, no. 1, p. 4307, 2018

    2018

  49. [57]

    Image-based profiling for drug discovery: due for a machine-learning up- grade?,

    S. N. Chandrasekaran, H. Ceulemans, J. D. Boyd, and A. E. Carpenter, “Image-based profiling for drug discovery: due for a machine-learning up- grade?,” Nature Reviews Drug Discovery, vol. 20, no. 2, pp. 145–159, 2021

    2021

  50. [58]

    Comprehensive transcription factor perturbations recapitulate fibroblast transcriptional states,

    K. M. Southard, R. C. Ardy, A. Tang, D. D. O’Sullivan, E. Metzner, K. Gu- ruvayurappan, and T. M. Norman, “Comprehensive transcription factor perturbations recapitulate fibroblast transcriptional states,” bioRxiv, 2024. 17

    2024

  51. [59]

    Pre- dicting cellular responses to novel drug perturbations at a single-cell res- olution,

    L. Hetzel, S. Boehm, N. Kilbertus, S. G¨ unnemann, F. Theis, et al. , “Pre- dicting cellular responses to novel drug perturbations at a single-cell res- olution,” Advances in Neural Information Processing Systems , vol. 35, pp. 26711–26722, 2022

    2022

  52. [60]

    Predicting transcriptional out- comes of novel multigene perturbations with gears,

    Y. Roohani, K. Huang, and J. Leskovec, “Predicting transcriptional out- comes of novel multigene perturbations with gears,” Nature Biotechnology, pp. 1–9, 2023

    2023

  53. [61]

    Y. Wu, Q. Liu, Y. Qiu, and L. Xie, “Deep learning prediction of chemical-induced dose-dependent and context-specific multiplex phenotype responses and its application to personalized alzheimer’s disease drug re- purposing,” PLOS Computational Biology, vol. 18, no. 8, p. e1010367, 2022

    2022

  54. [62]

    Dig- itize your biology! modeling multicellular systems through interpretable cell behavior,

    J. A. Johnson, G. L. Stein-O’Brien, M. Booth, R. Heiland, F. Kurtoglu, D. R. Bergman, E. Bucher, A. Deshpande, A. Forjaz, M. Getz, et al., “Dig- itize your biology! modeling multicellular systems through interpretable cell behavior,” bioRxiv, 2023

    2023

  55. [63]

    Tackling the challenges of new approach methods for predicting drug ef- fects from model systems,

    P. D. Pang, S. M. Ahmed, M. Nishiga, N. L. Stockbridge, and J. C. Wu, “Tackling the challenges of new approach methods for predicting drug ef- fects from model systems,” Nature Reviews Drug Discovery , 2024

    2024

  56. [64]

    Drug targeting in psychi- atric disorders—how to overcome the loss in translation?,

    K. Khodosevich, K. Dragicevic, and O. Howes, “Drug targeting in psychi- atric disorders—how to overcome the loss in translation?,” Nature Reviews Drug Discovery, vol. 23, no. 3, pp. 218–231, 2024

    2024

  57. [65]

    A deep learning framework for high-throughput mechanism-driven phenotype compound screening and its application to covid-19 drug repurposing,

    T.-H. Pham, Y. Qiu, J. Zeng, L. Xie, and P. Zhang, “A deep learning framework for high-throughput mechanism-driven phenotype compound screening and its application to covid-19 drug repurposing,” Nature ma- chine intelligence , vol. 3, no. 3, pp. 247–257, 2021

    2021

  58. [66]

    Integration of toxicogenomics and physiologically based pharmacokinetic modeling in human health risk as- sessment of perfluorooctane sulfonate,

    Q. Chen, W.-C. Chou, and Z. Lin, “Integration of toxicogenomics and physiologically based pharmacokinetic modeling in human health risk as- sessment of perfluorooctane sulfonate,” Environmental Science & Technol- ogy, vol. 56, no. 6, pp. 3623–3633, 2022

    2022

  59. [67]

    A context-aware deconfounding au- toencoder for robust prediction of personalized clinical drug response from cell-line compound screening,

    D. He, Q. Liu, Y. Wu, and L. Xie, “A context-aware deconfounding au- toencoder for robust prediction of personalized clinical drug response from cell-line compound screening,” Nature Machine Intelligence , vol. 4, no. 10, pp. 879–892, 2022

    2022

  60. [68]

    Towards out- of-distribution generalization: A survey,

    J. Liu, Z. Shen, Y. He, X. Zhang, R. Xu, H. Yu, and P. Cui, “Towards out- of-distribution generalization: A survey,” arXiv preprint arXiv:2108.13624, 2021

    2021 arXiv

  61. [69]

    Unlocking gene regulation with sequence-to-function models,

    A. Sasse, M. Chikina, and S. Mostafavi, “Unlocking gene regulation with sequence-to-function models,” Nature methods , vol. 21, no. 8, pp. 1374– 1377, 2024. 18

    2024

  62. [70]

    Towards the prediction of general biomolecular interactions with ai,

    M. Baek, “Towards the prediction of general biomolecular interactions with ai,” Nature Methods, vol. 21, no. 8, pp. 1382–1383, 2024

    2024

  63. [71]

    Progress and opportunities of foundation models in bioinformatics,

    Q. Li, Z. Hu, Y. Wang, L. Li, Y. Fan, I. King, G. Jia, S. Wang, L. Song, and Y. Li, “Progress and opportunities of foundation models in bioinformatics,” Briefings in Bioinformatics , vol. 25, no. 6, p. bbae548, 2024

    2024

  64. [72]

    A survey of uncertainty in deep neu- ral networks,

    J. Gawlikowski, C. R. N. Tassi, M. Ali, J. Lee, M. Humt, J. Feng, A. Kruspe, R. Triebel, P. Jung, R. Roscher,et al., “A survey of uncertainty in deep neu- ral networks,” Artificial Intelligence Review, vol. 56, no. Suppl 1, pp. 1513– 1589, 2023

    2023

  65. [73]

    Application of uncertainty quantification to artificial intelligence in healthcare: A review of last decade (2013–2023),

    S. Seoni, V. Jahmunah, M. Salvi, P. D. Barua, F. Molinari, and U. R. Acharya, “Application of uncertainty quantification to artificial intelligence in healthcare: A review of last decade (2013–2023),” Computers in Biology and Medicine, p. 107441, 2023

    2013

  66. [74]

    A survey on uncertainty quan- tification methods for deep learning,

    W. He, Z. Jiang, T. Xiao, Z. Xu, and Y. Li, “A survey on uncertainty quan- tification methods for deep learning,” arXiv preprint arXiv:2302.13425 , 2023

    2023

  67. [75]

    Explainable ai: A review of machine learning interpretability methods,

    P. Linardatos, V. Papastefanopoulos, and S. Kotsiantis, “Explainable ai: A review of machine learning interpretability methods,” Entropy, vol. 23, no. 1, p. 18, 2020

    2020

  68. [76]

    Causal machine learning for healthcare and precision medicine,

    P. Sanchez, J. P. Voisey, T. Xia, H. I. Watson, A. Q. O’Neil, and S. A. Tsaftaris, “Causal machine learning for healthcare and precision medicine,” Royal Society Open Science , vol. 9, no. 8, p. 220638, 2022

    2022

  69. [77]

    Causal inference in drug discovery and de- velopment,

    T. Michoel and J. D. Zhang, “Causal inference in drug discovery and de- velopment,” Drug discovery today , vol. 28, no. 10, p. 103737, 2023

    2023

  70. [78]

    A gentle introduction to conformal prediction and distribution-free uncertainty quantification,

    A. N. Angelopoulos and S. Bates, “A gentle introduction to conformal prediction and distribution-free uncertainty quantification,” arXiv preprint arXiv:2107.07511, 2021

    2021 arXiv

  71. [79]

    emosaic: Multi-modal out-of- distribution uncertainty quantification streamlines large-scale polypharma- cology,

    A. Badkul, L. Xie, S. Zhang, and L. Xie, “emosaic: Multi-modal out-of- distribution uncertainty quantification streamlines large-scale polypharma- cology,” bioRxiv, 2024

    2024

  72. [80]

    Appli- cations of genome-scale metabolic models to the study of human diseases: A systematic review,

    N. Cortese, A. Procopio, A. Merola, P. Zaffino, and C. Cosentino, “Appli- cations of genome-scale metabolic models to the study of human diseases: A systematic review,” Computer Methods and Programs in Biomedicine , p. 108397, 2024

    2024

  73. [81]

    Protocol for hybrid flux balance, statis- tical, and machine learning analysis of multi-omic data from the cyanobac- terium synechococcus sp. pcc 7002,

    S. Vijayakumar and C. Angione, “Protocol for hybrid flux balance, statis- tical, and machine learning analysis of multi-omic data from the cyanobac- terium synechococcus sp. pcc 7002,” STAR protocols , vol. 2, no. 4, p. 100837, 2021. 19

    2021

  74. [82]

    Integration of machine learning and genome- scale metabolic modeling identifies multi-omics biomarkers for radiation resistance,

    J. E. Lewis and M. L. Kemp, “Integration of machine learning and genome- scale metabolic modeling identifies multi-omics biomarkers for radiation resistance,” Nature Communications, vol. 12, no. 1, p. 2700, 2021

    2021

  75. [83]

    Unsupervised representa- tion learning on high-dimensional clinical data improves genomic discovery and prediction,

    T. Yun, J. Cosentino, B. Behsaz, Z. R. McCaw, D. Hill, R. Luben, D. Lai, J. Bates, H. Yang, T.-H. Schwantes-An, et al. , “Unsupervised representa- tion learning on high-dimensional clinical data improves genomic discovery and prediction,” Nature Genetics, vol. 56, no. 8, pp. 1...

    2024

  76. [84]

    Predicting pharmacodynamic effects through early drug discovery with artificial intelligence-physiologically based pharmacoki- netic (ai-pbpk) modelling,

    K. Wu, X. Li, Z. Zhou, Y. Zhao, M. Su, Z. Cheng, X. Wu, Z. Huang, X. Jin, J. Li, et al. , “Predicting pharmacodynamic effects through early drug discovery with artificial intelligence-physiologically based pharmacoki- netic (ai-pbpk) modelling,” Frontiers in Pharmacology, vol....

    2024

  77. [85]

    Whole person research and coordination center,

    National Center for Complementary and Integrative Health (NCCIH), “Whole person research and coordination center,” 2024. Accessed: 2024- 11-25. 20

    2024

Pith tools

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