WKWoo Youn Kim

Woo Youn Kim

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Woo Youn Kim has a total of 357 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Predicting conformational changes induced by biomolecular interactions

    by Yeongnam Bae, Jinheon Baek, Eugene Choi, Jeong Hun Choi, Hyeongyu Han, Hyosoon Jang, Woosung Jeon, Jaehyeong Jo, Dongki Kim, Hyeongwoo Kim, et al.

    Biomolecular interactions trigger structural changes of protein complexes essential for cellular signaling, regulation, and drug action. Modern co-folding models can now generate detailed structures of multi-component complexes directly from sequence. However, recent evaluations reveal that they frequently fail to capture specific conformational states induced by a ligand or partner. Here, we show that explicitly modeling the structural transition from an unbound to a bound state, rather than generating the bound complex alone, better resolves these interaction-induced states. Our generative learning framework, K-Fold, transforms independently predicted unbound component priors into an assembled complex through an ultrafast diffusion bridge, operating entirely without multiple sequence alignments. By eliminating co-evolutionary dependencies, K-Fold accurately recovers non-coevolved interfaces in antibody-antigen and targeted-degrader complexes, as well as ligand-directed regulatory states in GPCRs and kinases. Furthermore, by coupling this diffusion bridge with diverse monomeric equilibrium ensembles, K-Fold provides a computationally tractable foundation for sampling complex multi-state landscapes without retraining. These results demonstrate that formulating molecular assembly as an explicit structural transition moves structure prediction beyond static geometries toward capturing biomolecular dynamics.

  • Open Access

    Synthesis-aware generative design in trillion-scale chemical spaces for automated drug discovery

    by Seonghwan Seo, Yulseung Sung, Sang-Yeon Hwang, Mincheol Kang, Joonseong Lee, Samuele Bordi, Luka Raguž, Jihye Choi, Daniil Melnichenko, Wan Namkung, et al.

    Autonomous drug discovery via generative molecular design is critically bottlenecked by the production of chemically intractable structures. Multi-trillion-scale make-on-demand libraries guarantee synthetic feasibility, but conventional virtual screening cannot efficiently navigate these vast spaces or address multi-parameter optimization. We introduce Hyper Screening X, a generative AI framework integrating structure-based design with automated synthesis hardware. By encoding deterministic reaction logic in a generative flow network, the framework evaluates only 10 million candidates to navigate an 11-trillion-compound space, jointly optimizing physicochemical properties while ensuring compatibility with automated synthesis. Prospective validation against an SLC1A5 variant with a cryptic interface yielded a 96% synthesis success rate and a 50% functional hit rate, identifying two first-in-class lead compounds. This synthesis-aware strategy advances autonomous, closed-loop drug discovery.

  • Open Access

    On the limits and opportunities of AI reviewers: Reviewing the reviews of Nature-family papers with 45 expert scientists

    by Seungone Kim, Dongkeun Yoon, Kiril Gashteovski, Juyoung Suk, Jinheon Baek, Pranjal Aggarwal, Ian Wu, Viktor Zaverkin, Spase Petkoski, Daniel Schrider, et al.

    As AI capabilities advance, AI reviewers (agents that read a manuscript and produce a referee report) are beginning to be deployed in scientific peer review. Yet, their credibility remains in question: many scientists think that they cannot judge scientific validity, while others are more optimistic about their readiness. In this work, we conduct a study in which 45 scientists in Physical, Biological, and Health Sciences spend 469 hours rating 2,960 individual criticisms (each targeting one aspect of a paper) from human-written and AI-generated reviews of 82 Nature-family papers, scoring each on correctness, significance, and evidence sufficiency. On a composite rating of reviews across all three dimensions, an AI reviewer powered by GPT-5.2 scores above each paper's top-rated human reviewer (60.0% vs 48.2% of criticisms, p=0.009). Also, AI reviewers' correct criticisms are more often rated significant, and surface a distinct 26% of legitimate issues that no human raises. However, AI reviewers are more often incorrect, overlap with each other far more than humans do with each other, and exhibit 16 weaknesses rare in human reviews. Overall, our results position current AI reviewers as complements to, not substitutes for, human reviewers.


  • Open Access

    How Do Co-folding Models Organize Structural Information?

    by Minha Park, Shinwoo Kim, Seokhyun Moon, Hyeongwoo Kim, Gyuho Jeon, Woo Youn Kim

    Co-folding models accurately predict biomolecular complexes, but how their internal representations support joint structure prediction remains unclear. We analyze the Boltz-1 trunk by decomposing its pair representation into intra-chain and inter-chain blocks and applying ablation, layer-wise representation geometry, and activation patching. Our results show that structural information is distributed across three streams: the single representation carries token-internal atomic detail, the intra-chain pair representation encodes chain-level geometry, and the inter-chain pair representation controls relative molecular arrangement. Across the trunk depth, representations follow a Mix–Compress–Refine trajectory, while linear probes reveal that intra-chain distance content is largely supplied by MSA preconditioning and inter-chain distance content is constructed progressively across Pairformer layers at each recycle. Apo–holo patching further shows that ligand-induced backbone changes are not fully resolved before diffusion; instead, the diffusion module reconciles partially inconsistent intra- and inter-chain constraints. These findings provide a mechanistic view of how co-folding trunks organize structural information and suggest more partition-aware designs for future biomolecular structure prediction models.

  • Open Access

    AtlasFold: Protein structure prediction with metagenomic-scale language models

    by Seonghwan Seo, Hyeongwoo Kim, Seokhyun Moon, Woo Youn Kim, Team Kaist

    Protein language models (PLMs) trained on evolutionary sequences learn representations that encode protein structure, enabling direct structure prediction without multiple-sequence alignments (MSAs). Here we present the Atlas model family, an open and trainable system spanning protein language modeling, monomer folding, and protein-complex prediction. AtlasLM-3B is a 3B-scale language model trained with masked language modeling on approximately 1.56 billion sequences, including metagenomic data, and outperforms the similarly sized ESM2-3B in unsupervised contact prediction. Building on these representations, AtlasFold predicts all-atom protein structures and achieves state-of-the-art accuracy among PLM-based folding models. Fine-tuning AtlasFold for protein-complex prediction produces AtlasFold-Multimer (AtlasFold-M), whose antibody–antigen prediction performance is comparable to that of AlphaFold3 and ESMFold2. This protein-specific folding architecture enables fast, memory-efficient inference with AtlasFold and AtlasFold-M. By releasing the training code and data, stage checkpoints, and model weights under the MIT License, we provide a foundation for advancing PLM-based protein structure prediction. 1

  • Open Access

    Autonomous bioisosteric replacement for multi-property optimization in drug design

    by Hyeongwoo Kim, Seokhyun Moon, Wonho Zhung, Shinwoo Kim, Jaechang Lim, Woo Youn Kim

    Optimizing molecular properties while preserving biological activity is a central challenge in drug design. Bioisosteric replacement, which substitutes a molecular fragment with a chemically or biologically analogous moiety, offers a powerful strategy for fine-tuning properties without disrupting target binding. However, existing in silico approaches often rely on expert-defined modification sites or struggle to modulate multiple molecular properties simultaneously. Here, we present DeepBioisostere, a deep generative model that performs end-to-end bioisosteric replacement by autonomously selecting and substituting molecular fragments to satisfy multiple target properties. The model captures complex relationships across the molecular graph, enabling the optimization of sophisticated properties such as drug-likeness and synthetic accessibility. By learning from experimental bioassay data, DeepBioisostere proposes replacements that maintain biological activities, even generating potential bioisosteres beyond the training data. We demonstrate the effectiveness of the model in computational hit-to-lead optimization scenarios, highlighting its potential to accelerate rational molecular design without relying on expert heuristics or pre-established substitution rules.

  • Open Access

    A Comprehensive Dataset of Chemical Reactions Covering Second and Third Row Elements with Million-Scale Quantum Chemical Calculations

    by Kyunghoon Lee, Sungwoo Kang, Beomgyu Kang, Won-Joon Son, Seungmin Lee, Dae Sin Kim, Woo Youn Kim

    We present an extended quantum chemical reaction database, called Reaction-QM, containing millions of reactions with fully optimized reactant, product, and transition state (TS) structures. The database comprises three subsets: The first set contains 2,304,303 organic reactions with up to ten heavy atoms and various elements (H, C, O, N, F, Si, P, S, and Cl) characterized at the GFN2-xTB level of theory, expanding the element scope and the size of the dataset. The second set consists of 199,890 reactions characterized at the B3LYP-D3/TZVP level of theory. All TS structures were validated through vibrational analysis and intrinsic reaction coordinate (IRC) calculations, ensuring the correct assignment of reactants and products. The final set contains full IRC trajectories at the DFT level for all 199,890 reactions, consisting of total 22,984,707 molecular configurations, offering a rich dataset for training machine learning interatomic potentials (MLIPs). Together, the databases can serve as a new data source for developing and benchmarking ML-based models for reaction modeling tasks, including TS structure prediction, activation barrier estimation, and the construction of reactive MLIPs.


  • Open Access

    Harnessing generative AI for efficient organic materials discovery in low-data regimes

    by Jun Hyeong Kim, Kyunghoon Lee, Hyeonsu Kim, MinSoo Kang, Suk-Ku Chang, Yinglan Jin, Dongwook Kim, Woo Youn Kim

    Generative AI has emerged as a powerful tool for the discovery of organic light-emitting diode (OLED) materials. However, its practical application remains underexplored due to the small datasets and difficulties in ensuring molecular synthesizability. To overcome these challenges, we introduce a building block-based autoregressive generative model. Trained on a dataset of approximately 1000 OLED molecules, the model demonstrated refined control over key thermally activated delayed fluorescence (TADF) properties, including S1 energy and the singlet–triplet energy gap ΔEST, while generating structurally novel candidates through strategic repurposing of building blocks not previously associated with TADF activity. In addition, we experimentally validated its potential by synthesizing four AI-designed green emitters and integrating them into OLED devices, achieving external quantum efficiencies of up to 11.22% at 1000 cd m−2. It achieved more than a 100-fold reduction in the computational cost of quantum chemical calculations compared to conventional heuristic methods. This work bridges the gap between generative molecular design and experimental realization, showcasing a pathway to overcome data scarcity and unlock innovative discovery of optoelectronic materials.


  • Open Access

    Synthesis of 3-desoxycollinoketone B and its ability to reduce Alzheimer-associated misfolded proteins

    by Seongrok Heo, Minhae Cha, Wonho Zhung, JiMin Kim, Sirisuk Keereewan, Illhwan Cho, MinSeol Park, Wonbin Seo, Heewon Shin, Soljee Yoon, et al.

    Collinolactone, featuring a 7/10/6 tricyclic core, has been proposed to be biosynthesized via a transannular [6 + 4] cycloaddition reaction. Besides its intriguing architecture, collinolactone holds pharmaceutical promises due to its ability to disrupt amyloid-β (Aβ) and tau aggregation, which are specifically found as disease culprits in the brains of Alzheimer’s disease (AD) patients and are key targets in current drug discovery efforts. However, challenges associated with its acquisition from a natural source and limited pharmacokinetic properties have hampered its further studies. Herein, we report the design, synthesis, and biological evaluation of 3-desoxycollinoketone B, a collinolactone derivative with improved pharmacokinetics for AD treatment. A stereoselective transannular [6 + 4] cycloaddition efficiently constructs the tricyclic core, allowing its scalable synthesis. AI-assisted binding prediction and simulations not only indicate superior binding of 3-desoxycollinoketone B to Aβ and tau aggregates to collinolactone, but also suggest a mechanistic basis for fibril destabilization. In vitro studies confirm its inhibition and dissociation of Aβ and tau fibrils, while in vivo experiments in AD mouse models show substantial amelioration of cognitive functions and Aβ/tau-associated pathology.

  • Riemannian denoising model for molecular structure optimization with chemical accuracy

    by Jeheon Woo, Seonghwan Kim, Jun Hyeong Kim, Woo Youn Kim

    Here we introduce a framework for molecular structure optimization using a denoising model on a physics-informed Riemannian manifold (R-DM). Unlike conventional approaches operating in Euclidean space, our method leverages a Riemannian metric that better aligns with molecular energy change, enabling more robust modeling of potential energy surfaces. By incorporating internal coordinates reflective of energetic properties, R-DM achieves chemical accuracy with an energy error below 1 kcal mol−1. Comparative evaluations on QM9, QM7-X and GEOM datasets demonstrate improvements in both structural and energetic accuracy, surpassing conventional Euclidean-based denoising models. This approach highlights the potential of physics-informed coordinates for tackling complex molecular optimization problems, with implications for tasks in computational chemistry and materials science.