Computational biology methods
Yehui Yang, Zelin Zang, Changxi Chi et al.
This addresses the challenge of generalizing to out-of-distribution cell states in single-cell annotation, offering a robust solution for biological research.
Tianyu Liu, Weihao Xuan, Hao Wu et al.
This addresses the challenge of building reliable AI copilots for real-world pathology scenarios, though it appears incremental as it builds on existing multimodal models with reinforcement learning enhancements.
Chaoran Cheng, Jiaqi Guan, Milong Ren et al.
This work provides a new versatile foundation for all-atom generative modeling in protein design, enabling seamless adaptation to non-canonical amino acids for the first time.
Yutang Ge, Guojiang Zhao, Sihang Li et al.
For protein engineers, this provides a data-efficient framework to translate natural language requirements into viable protein sequences.
Chuang Zhao, Hongke Zhao, Xiaofang Zhou et al.
This work improves clinical reasoning for healthcare applications by enabling models to better internalize complex case nuances, though it appears incremental as it builds on existing test-time training and calibration methods.
Yixian Xu, Yusong Wang, Shengjie Luo et al.
For generative modeling of symmetric data like molecular structures, this provides a principled framework that simplifies learning and guarantees correct sampling, outperforming heuristic alignment methods.
Montgomery Bohde, Hongxuan Liu, Mrunali Manjrekar et al.
This work addresses the problem of de novo structural elucidation from mass spectra, a key bottleneck in analytical chemistry, with a scalable approach that significantly outperforms existing methods.
Junda Ying, Yuxuan Wang, Bowen Yang et al.
For computational biologists studying cellular differentiation and lineage branching, USB provides a rigorous microscopic interpretation of birth-death events, addressing a key limitation of existing continuous optimal transport methods.
Han Zhang, Guo-Hua Yuan, Chaohao Yuan et al.
This work provides a generative cellular world model for in silico simulation of cell states and perturbation responses, addressing the need for virtual cells in biological discovery and perturbation screening.
Zhenyu Wang, Geyan Ye, Wei Liu et al.
This work addresses the need for more reliable and interpretable virtual cell perturbation predictions for biological mechanism studies, representing a domain-specific advancement.
Andrew Shen, Shaul Druckmann, James Zou
For AI-driven scientific discovery, AR provides a method to overcome LLM mode collapse, enabling more creative and effective solution generation in biomedicine.
Dongxin Ye, Fang Hu, Han Hu et al.
This benchmark provides a standardized evaluation framework for NFMs in viral genomics, addressing both biological understanding and biosecurity risks, which is critical for the biomedical community.
Edward De Brouwer, Carl Edwards, Alexander Wu et al.
Provides the first standard benchmark for in silico phenotypic screening, a key capability for virtual cell models in drug discovery.
Jueon Park, Wonjune Jang, Chanhwi Kim et al.
This work addresses the need for reliable toxicity modeling in drug discovery and safety assessment, though it is incremental as it builds on existing benchmarks and methods.
Yucheng Xing, Pei Liu, Jingying Ma et al.
For computational pathology, MIST addresses the modality bottleneck in MIL by incorporating molecular information without requiring transcriptomics at inference, yielding consistent improvements across diverse endpoints.
Zhe Zhang, Yuanning Feng, Yuxuan Song et al.
For researchers and practitioners in protein design and virtual screening, DCFold dramatically reduces inference time while maintaining state-of-the-art accuracy, enabling practical deployment of high-quality protein structure generation.
Xinrui Chen, Yizhen Luo, Siqi Fan et al.
This work addresses the challenge of designing functional proteins without evolutionary templates, benefiting biotechnology and medicine by improving both functionality and foldability.
Xiaoqing Lian, Pengsen Ma, Tengfeng Ma et al.
This work addresses the trade-off between scalability and biological depth in drug discovery, offering a novel framework for phenotypic virtual screening.
Dejun Lin, Simon Chu, Vishanth Iyer et al.
This addresses the problem of scaling biomolecular modeling for researchers by providing a scalable pathway to model massive systems, representing an incremental improvement in computational efficiency.
Jianshen Zhu, Raveena Rai, Taiyo Sohkawa et al.
This work provides a tractable method for exact inverse design of copolymers, addressing a gap in polymer informatics for practitioners needing optimal molecular structures.