FYFan Yang

Fan Yang

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

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Fan Yang has a total of 82 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    UniPert-G2CP bridges genetic and chemical screens from molecular representation to phenotype modeling

    by Yiming Li, Min Zeng, Jun Zhu, Linjing Liu, Fang Wang, Longkai Huang, Fan Yang, Min Li, Jianhua Yao


  • Open Access

    Functional protein design and enhancement with ontology reinforcement iteration

    by Bing He, Chenchen Qin, Yu Zhao, Long-Kai Huang, Zihan Wu, Fang Wang, Fandi Wu, Fan Yang, Jianhua Yao

    The discrepancy between computational modeling and experimental performance remains a major challenge in protein engineering. We present ORI (Ontology Reinforcement Iteration), a scalable framework integrating ontology-conditioned decoding with reinforcement learning from experimental feedback (RLWF). ORI leverages structured ontologies as semantic prompts to impose multi-level constraints, enabling controllable and interpretable protein generation. A closed-loop iterative workflow—comprising generation, experimental measurement, and model updating—enables continuous optimization under real-world objectives. We demonstrate ORI’s practical applicability through diverse tasks, including enzymatic activity optimization, thermal stability enhancement, and multifunctional protein engineering. Using this framework, we engineer variants with substantial improvements over natural baselines, such as a lysozyme with 100-fold higher activity, a chitinase stable at 85 °C, and dual-function enzymes exhibiting both lysozyme and chitinase activities. These results establish ORI as a robust technical platform for efficient, multi-objective protein engineering in real-world experimental settings.


  • Open Access

    MolSpecFlow: Mass-Constrained Hybrid Flow Matching for Joint Molecular-Spectral Analysis

    by Yu Wang, Fan Yang, Kaikun Xu, Li Yuan, Jun Zhu, Jingjie Zhang, Zhenchao Tang, Yatao Bian, Cheng Chang, Yonghong Tian, et al.

    Identifying the “dark matter” of the chemical universe requires bridging the fundamental gap between molecules and mass spectra. Existing approaches struggle to accurately map between these distinct data formats, often yielding results that are either chemically implausible or inconsistent with the physical evidence. We introduce MolSpecFlow, a unified foundation model pretrained on 100 million molecules and 42 million spectra, which leverages a hybrid flow matching framework to orchestrate optimal transport paths for spectral peaks and discrete probability flows for molecular tokens. Furthermore, to guarantee the physicochemical validity of the generated structures, we incorporate explicit rule-based constraints into the generative process, utilizing a token-level mass control mechanism that strictly enforces alignment with the precursor molecular weight. MolSpecFlow establishes a new state-of-the-art on the MassSpecGym benchmark, surpassing diffusion and retrieval baselines in de novo generation (Top-1 accuracy +35%), spectral simulation, and molecular retrieval, demonstrating the power of physics-grounded unified modeling.

  • Open Access

    De novo design of epitope-specific antibodies via a structure-driven computational workflow

    by Fandi Wu, Yu Zhao, JiaXiang Wu, Biaobin Jiang, Bing He, Longkai Huang, Chenchen Qin, Yang Xiao, Fan Yang, Rubo Wang, et al.

    Accurate modeling of antibody-antigen complex structures holds significant potential for advancing biomedical research and the design of therapeutic antibodies. Compared to general proteins, progress in antibody structure prediction and design has been slow, and antibody discovery is still based on time-consuming animal immunization or library screening methods. Here, we present tFold System, a high-throughput computational workflow that integrates antibody structure prediction (tFold-Ab), antibody-antigen complex modeling (tFold-Ag), structure-guided virtual screening, and de novo epitope-specific antibody design. Using this system, we de novo design monoclonal antibodies (mAbs) against four therapeutically relevant antigens: influenza hemagglutinin (Flu A), PD-1, PD-L1, and SARS-CoV-2 RBD (SC2RBD). Experimental validation by surface plasmon resonance (SPR) following high-throughput screening via phage display shows the designed antibodies achieve nanomolar binding affinities and precise epitope targeting, demonstrating the efficiency of the integrated computational-experimental pipeline. Our results demonstrate that tFold System overcomes key limitations of existing methods by enabling rapid, high-throughput antibody discovery against user-defined epitopes.


  • A pre-trained large generative model for translating single-cell transcriptomes to proteomes

    by Linjing Liu, Wei Li, Fang Wang, Yiming Li, Long-Kai Huang, Ka-Chun Wong, Fan Yang, Jianhua Yao

    Measuring protein abundance at the single-cell level can facilitate a high-resolution understanding of biological mechanisms in cellular processes and disease progression. However, current single-cell proteomic technologies face challenges such as limited coverage, constrained throughput and sensitivity, batch effects, high costs and stringent experimental operations. Inspired by the translation procedure in both natural language processing and the genetic central dogma, we propose a pre-trained, large generative model named single-cell translator (scTranslator). scTranslator can generate multi-omics data by inferring the missing single-cell proteome based on the transcriptome. Through systematic benchmarking and validation on independent datasets, we have confirmed the accuracy, stability and flexibility of scTranslator across various profiling techniques (for example, CITE-seq, spatial CITE-seq, REAP-seq, NEAT-seq), cell types (for example, monocytes, macrophages, T cells, B cells), tissues (for example, blood, lung, brain) and a wide range of disease contexts, including infectious, metabolic and oncologic conditions. Furthermore, scTranslator shows its superiority in assisting various downstream analyses and applications, including gene/protein interaction inference, perturbation prediction, cell clustering, batch correction and cell origin recognition in pan-cancer data.


  • Open Access

    Genetic-to-Chemical Perturbation Transfer Learning Through Unified Multimodal Molecular Representations

    by Yiming Li, Min Zeng, Jun Zhu, Linjing Liu, Fang Wang, Longkai Huang, Fan Yang, Min Li, Jianhua Yao

    Artificial Intelligence virtual cell (AIVC) holds transformative potential for biomedical research. Central to this vision is the systematic modeling of genetic and chemical perturbation phenotypes to accurately predict cellular dynamic states from diverse interventions. However, disparities in screening agents, library scales, experimental technologies, and data production efficiency hinder the integration, modeling, and analysis of the cross-data. Here we present UniPert-G2CP , a two-phase deep learning approach comprising i) UniPert, a multimodal molecular representation model that bridges genetic and chemical domains, and ii) G2CP ( Genetic-to-Chemical Perturbation transfer learning), which systematically transforms CRISPR screen-based genetic insights into chemical perturbation modeling for cost-effective in silico drug screening. UniPert not only encodes multimodal perturbagens into a unified functionally interpretable sematic embedding space, but also improves phenotypic effect prediction for previously unseen gene perturbations and drug treatments. Building upon UniPert, G2CP successfully modeled large-scale cellular post-perturbation states spanning 4,994 gene and 7,821 compound perturbagens, while reducing modeling data costs by over 60%. We demonstrate that UniPert-G2CP enables efficient, generalizable simulations of multicellular, multi-domain perturbation cause-effect spaces, revealing differential cellular biological causality and informing mechanism-driven therapy. UniPert-G2CP opens new avenues for biological causal foundation model building, AIVC creation, and AI-powered precision medicine.

  • Open Access

    CD-GPT As a Biological Foundation Model Bridging the Gap between Molecular Sequences Through Central Dogma

    by Xiao Zhu, Chenchen Qin, Fang Wang, Fan Yang, Bing He, Yu Zhao, Jianhua Yao

    The central dogma serves as a fundamental framework for understanding the flow and expression of genetic information within living organisms, facilitating the connection of diverse biological sequences across molecule types. In this study, we present CD-GPT (Central Dogma Generative Pretrained Transformer), a generative biological foundation model with 1 billion parameters, aiming to capture the sequence relationships between DNA, RNA, and proteins. We model sequences in a unified representational space and employ a shared, multi-molecule vocabulary to narrow their distances in the embedding space effectively. Through extensive pretraining on nucleotide and amino acid sequence data, CD-GPT exhibits exceptional performance in a wide range of predictive and generative downstream tasks, including mono-molecular and multi-molecular analyses. Notably, CD-GPT excels in tasks such as genomic element detection, protein property prediction, RNA-protein interaction identification and also generative tasks like protein generation and reverse translation. The versatility of CD-GPT opens up promising avenues for advanced multi-omics analysis.


  • Open Access

    A pre-trained large generative model for translating single-cell transcriptome to proteome

    by Linjing Liu, Wei Li, Ka-Chun Wong, Fan Yang, Jianhua Yao

    Proteins are crucial for life, and measuring their abundance at the single-cell level can facilitate a high-resolution understanding of biological mechanisms in cellular processes and disease progression. However, current single-cell proteomic technologies face challenges such as limited coverage, throughput, and sensitivity, as well as batch effects, high costs, and stringent experimental operations. Drawing inspiration from the translation procedure of both natural language processing (NLP) and the genetic central dogma, we propose a pre-trained, large generative model named scTranslator (single-cell translator). scTranslator is align-free and capable of generating multi-omics data by inferring the missing single-cell proteome based on the transcriptome. Systematic benchmarking confirms the accuracy, stability, and flexibility of scTranslator across various quantification techniques, cell types, and conditions. Furthermore, scTranslator has demonstrated its superiority in assisting various downstream analyses and applications, including gene/protein interaction inference, gene pseudo-knockout, cell clustering, batch correction, and cell origin recognition on pan-cancer data.


  • Open Access

    Deep Domain Adversarial Neural Network for the Deconvolution of Cell Type Mixtures in Tissue Proteome Profiling

    by Fang Wang, Fan Yang, Long-Kai Huang, Jiangning Song, Robin B. Gasser, Ruedi Aebersold, Guohua Wang, Jianhua Yao, Jiang Qian

    Cell type deconvolution is a computational method for the determination/resolution of cell type proportions from bulk sequencing data, frequently used for the analysis of divergent cell types in tumor tissue samples. However, deconvolution technology is still in its infancy for the analysis of cell types using proteomic data due to challenges with repeatability/reproducibility, variable reference standards and the lack of single-cell proteomic reference data. Here, we developed a novel deep learning-based deconvolution method (scpDeconv) tailored to proteomic data. scpDeconv uses an autoencoder to leverage the information from bulk proteomic data to improve the quality of single-cell proteomic data, and employs a domain adversarial architecture to bridge the single-cell and bulk data distributions and transfer labels from single-cell data to bulk data. Extensive experiments validated the performance of scpDeconv in the deconvolution of proteomic data produced from various species/sources and different proteomic technologies. This method should find broad applicability to areas including tumor microenvironment interpretation and clinical diagnosis/classification.


  • Open Access

    Preoperative Prediction of Lymph Node Metastasis in Colorectal Cancer with Deep Learning

    by Hailing Liu, Yu Zhao, Fan Yang, Xiaoying Lou, Feng Wu, Hang Li, Xiaohan Xing, Tingying Peng, Bjoern Menze, Junzhou Huang, et al.

    Objective . To develop an artificial intelligence method predicting lymph node metastasis (LNM) for patients with colorectal cancer (CRC). Impact Statement . A novel interpretable multimodal AI-based method to predict LNM for CRC patients by integrating information of pathological images and serum tumor-specific biomarkers. Introduction . Preoperative diagnosis of LNM is essential in treatment planning for CRC patients. Existing radiology imaging and genomic tests approaches are either unreliable or too costly. Methods . A total of 1338 patients were recruited, where 1128 patients from one centre were included as the discovery cohort and 210 patients from other two centres were involved as the external validation cohort. We developed a Multimodal Multiple Instance Learning (MMIL) model to learn latent features from pathological images and then jointly integrated the clinical biomarker features for predicting LNM status. The heatmaps of the obtained MMIL model were generated for model interpretation. Results . The MMIL model outperformed preoperative radiology-imaging diagnosis and yielded high area under the curve (AUCs) of 0.926, 0.878, 0.809, and 0.857 for patients with stage T1, T2, T3, and T4 CRC, on the discovery cohort. On the external cohort, it obtained AUCs of 0.855, 0.832, 0.691, and 0.792, respectively (T1-T4), which indicates its prediction accuracy and potential adaptability among multiple centres. Conclusion . The MMIL model showed the potential in the early diagnosis of LNM by referring to pathological images and tumor-specific biomarkers, which is easily accessed in different institutes. We revealed the histomorphologic features determining the LNM prediction indicating the model ability to learn informative latent features.