Aug 2026· PLoS Computational Biology· Vol 22, pp. e1014499· 0 citations· 46 references
Medicine
Abstract
CLDN18.2 is a promising tumor-specific antigen; however, the development of therapeutic antibodies against it is challenged by the need for simultaneous optimization of affinity and developability. To address this, we present cdrGPT, a deep learning framework based on GPT-2 for de novo generation of complementarity-determining region H3 (CDRH3) sequences. Our approach integrates pre-training on the Observed Antibody Space (OAS) database with structural templating derived from the known antibody zolbetuximab. Generated sequences were iteratively refined through rejection sampling and fine-tuned against a multi-parameter objective function encompassing predicted affinity and MHC class II binding risk. From an initial set of 50,000 sequences, this screening pipeline yielded 313 high-confidence candidates. Subsequent analysis using evolutionary scale modeling 2 (ESM2) embeddings, principal component analysis (PCA), and clustering revealed three structurally distinct clusters, with intra-cluster cosine similarities exceeding 0.99. Validation of seven representative sequences from the dominant cluster using AlphaFold3 confirmed high structural fidelity to the zolbetuximab template, demonstrating a root mean square deviation (RMSD) of 1.331 Å for the CDRH3 loop and positional deviations of less than 0.4 Å for key paratope residues. These results indicate that the designed variants preserve the core binding mode of the parent antibody. This study establishes a feasible pipeline for integrating AI-generated CDRH3 loops into functional antibody scaffolds, providing a foundation for the accelerated development of therapeutics targeting CLDN18.2 and other clinically relevant antigens.
This work designs a biological prior-guided feature fusion framework that integrates pseudo-structural epitope knowledge and CDR-specific attention mechanisms via a mixture-of-experts architecture to effectively capture complex binding landscapes in antibody screening and drug residence time analysis.
G. Luo, Junkai Wang, Sizhe Zhang et al.· Bioinformatics· 0 citations
Predicting immunogenicity remains a critical challenge in vaccine design, autoimmunity treatment, and pharmaceutical development. Current AI tools rely on limited data inputs, typically only class-I MHC-peptide sequences, missing crucial structural and biochemical information. We developed the ImmunoFoundation Model (IFM), a multimodal deep learning system that integrates not only peptide sequences, 3D molecular structures, and biochemical properties but also TCR-MHC-peptide (class-II and class-I) to achieve superior immunogenicity prediction and enable peptide optimization for therapeutic applications.
IFM comprises three modules: (1) ESM3 transformer for embedding antigen-peptide, MHC, and TCR sequences from vast biomedical data; (2) geometric scattering transformer networks to capture molecular structure from AlphaFold3-predicted peptide-MHC complexes; (3) Autoencoder for biochemical property embedding including surface area and thermal stability. Cross-modal attention layers integrate these representations. Training utilized IEDB, VDJdb, McPAS-TCR, and TCR3d datasets totaling >300,000 samples across MHC class I and II.
The preliminary model achieved state-of-the-art performance on CEDAR cancer neoepitope datasets. Attention mechanism analysis revealed structural motifs influencing immunogenicity, distinguishing between KRAS G12V and G12D mutants. The model successfully predicted vaccine cassette immunogenicity and identified key peptide-MHC interaction sites. Current IFM development shows improved multimodal integration with enhanced predictive accuracy across viral and cancer peptide immunogenicity tasks.
IFM represents a paradigm shift in immunogenicity prediction by comprehensively modeling the complex antigen-MHC-TCR interaction system. Its generative capabilities enable peptide optimization for cancer vaccines and personalized immunotherapy, with potential applications in autoimmunity treatment and biologics development.
Yale Colton Center for Autoimmunity
Computational and Systems Immunology (COMP)
Smita Krishnaswamy, J. Rocha, Hiren Madhu et al.· Journal of Immunology· 0 citations
This thesis examines the integration of machine learning into computational structural biology, with an emphasis on modelling and predicting antibody–antigen interactions. Such interactions are fundamental to numerous biological processes and are central to therapeutic antibody design. Despite recent advances in AI-based protein structure prediction, antibodies remain particularly challenging targets due to the high variability of their complementarity-determining regions, the limited availability of experimental structures, and the lack of strong co-evolutionary signal.
To address these challenges, this work introduces several methodological contributions. In Chapter 2,DeepRank-GNN-esm incorporates embeddings from protein language models to replace computationally expensive evolutionary features, thereby improving both predictive performance and efficiency in scoring protein–protein complexes. In Chapter 3, a modelling pipeline is introduced that employs a flow-matching algorithm to effectively sample the conformational diversity of the antibody CDR-H3 loop. When integrated with ensemble docking, this approach significantly improves the accuracy of antibody–antigen complex modelling compared to existing methods. In Chapter 4, the thesis presents AbTune, a sequence-specific fine-tuning strategy for protein language models that enhances predictive performance across multiple antibody-related tasks, including structure prediction, mutation effect estimation, and binding affinity prediction, while remaining computationally efficient. In Chapter 5, DeepRank-Ab is developed as a geometric deep learning-based scoring function tailored to antibody–antigen complexes, achieving state-of-the-art performance in ranking near-native docking conformations. Chapter 6 summarizes the main findings of the thesis and discusses future research directions.
Collectively, these contributions demonstrate how machine learning can be applied to address key limitations in antibody modelling and to facilitate the rational design of antibody-based therapeutics.
Sequence-based protein druggability classification can support early target triage when structural information is unavailable, uncertain, or inconsistently linked to druggability labels. We present DrugPLMFormer, a sequence-first retrospective screening framework that combines frozen protein language model embeddings with self-attentive BiLSTM encoding, Transformer-based long-range modeling, optional physicochemical feature fusion, and compute-budgeted BO–CTCM model selection. Hyperparameters were selected through multi-fidelity screening within an approximately 200-evaluation budget, using a validation objective that combined AUPRC and MCC to balance threshold-free discrimination with operating-point stability, rather than to imply unrestricted generalization. On ProTar-II, using a 50% sequence-identity homology-aware split, DrugPLMFormer achieved 95.98% accuracy, 96.01% F1-score, 96.42% sensitivity, 95.61% specificity, and 0.981 ROC-AUC. Without using external data for training, tuning, threshold selection, or early stopping, the selected model showed favorable held-out mean performance on ProTar-II-Ind (96.62% accuracy, 0.9688 ROC-AUC) and DPI_CDF (96.20% accuracy, 0.9696 ROC-AUC). Paired external analyses indicated that accuracy and F1-score differences were numerically favorable but not statistically significant, whereas the ROC-AUC improvement on DPI_CDF was statistically supported. Train-to-external homology analysis showed that most external proteins had less than 50% sequence identity to the training set, although residual dataset shift and label heterogeneity may still affect generalization. With cached PLM embeddings, downstream CPU inference required approximately 1.0–1.2 ms per sequence, excluding tokenization and ESM-2 embedding generation. Overall, DrugPLMFormer provides a reproducible, leakage-aware framework for retrospective sequence-based druggability screening and target prioritization, while prospective validation and experimental confirmation remain necessary before operational deployment.
Zahra Kafi, Khosro Rezaee, Hossein Eslami· Journal of King Saud Univers...· 0 citations
Cyclin-dependent kinase 8 (CDK8) has emerged as a promising therapeutic target for acute myeloid leukaemia (AML). We developed an integrated computational pipeline combining neural network-based potency prediction with molecular dynamics (MD) simulations for CDK8 inhibitor discovery. A curated dataset of 1,200 unique CDK8 inhibitors was assembled from ChEMBL. The optimal neural network architecture (two hidden layers, 512→128 units) with dropout regularization achieved a test set r2 of 0.47 and RMSE of 0.87. Y-randomization testing confirmed genuine structure-activity relationships. Using an extrapolation-focused training strategy combined with a SELFIES-based genetic recombination algorithm, we generated novel molecular structures beyond the training distribution. SHAP analysis revealed fingerprint bits as critical determinants of potency. Applicability‑domain analysis confirmed that the novel hit falls within validated chemical space. The identified candidate exhibited a predicted pIC50 of 11.97, substantially exceeding the most potent training compound (pIC50 = 10.09) and compound 12 (pIC50 = 7.47). Molecular docking revealed Moldock scores of -143.9 kcal/mol for the novel hit versus -113.9 kcal/mol for compound 12. MD simulations demonstrated stable binding with the novel hit forming a highly stable hydrogen bond with Asp98. MM-GBSA calculations showed superior binding free energy for the novel hit (-99.10 vs. -47.77 kcal/mol).
B. R. Awad, M. Sargolzaei, H. Nikoofard· SAR and QSAR in environmenta...· 0 citations
DeepCGASPred is a cyclic GMP-AMP synthase (cGAS)-specific deep learning scoring function that integrates three-dimensional convolutional neural networks with multi-head attention mechanisms and composite structural descriptors, including Structural Protein–Ligand Interaction Fingerprints (SPLIF), hydrogen bond features, and extended connectivity fingerprints (ECFP).
Muhammad Junaid, Muhammad Zeeshan, Abbas Khan et al.· Discover Chemistry· 0 citations