{
  "id": 18200035,
  "title": "CARTnet: Template-guided B-cell Epitope Prediction with Contact-aware Attention",
  "url": "https://urgent.news/2026/11/04/cartnet-template-guided-b-cell-epitope-prediction-with-contact-aware",
  "topic": "science",
  "section": "Science",
  "published": "2026-11-04T00:00:00.000Z",
  "source": {
    "name": "bioRxiv",
    "slug": "biorxiv",
    "url": "https://www.biorxiv.org/content/10.64898/2026.11.03.764112v1?rss=1"
  },
  "original_language": "en",
  "account": "B-cells' ability to generate epitopes that recognize antigens remains a formidable challenge due to limited experimental data and a lack of comprehension regarding antibody-antigen interactions. The groundbreaking CARTnet framework, introduced in this study, addresses this issue by leveraging multimodal features and template-derived epitope information within a deep learning architecture. When evaluated against the latest state-of-the-art methods, CARTnet proved to be more effective across various assessment metrics on both internal and external datasets. Specifically, CARTnet achieved an Area Under the Precision-Recall Curve (AUPRC) of 0.243 on the internal test set, surpassing DiscoTope-3 (0.220), BepiPred-2 (0.140), and BepiPred-3 (0.198). This advantage was also evident on the external test set, where CARTnet attained an AUPRC of 0.267, outperforming DiscoTope-3 (0.192), BepiPred-2 (0.111), and BepiPred-3 (0.215). CARTnet's adaptability is further exemplified by its template-guided prediction capability, which allows limited prior epitope information to enhance the identification of additional epitope residues. Notably, incorporating a single template residue improved the AUPRC of the template-based variant from 0.225 to 0.403 on the internal test set. Moreover, the study revealed that epitope information derived from closely related viruses could be transferred across the Alphavirus genus, thereby augmenting epitope prediction accuracy. These findings establish CARTnet as a robust and precise tool for both de novo and template-guided B-cell epitope prediction.",
  "summary": "B-cell epitope prediction remains challenging due to the limited availability of experimentally characterized data and the incomplete understanding of antibody-antigen recognition. Here, we present CARTnet, a novel contact-aware deep learning framework that integrates multimodal features and template-derived epitope information. CARTnet demonstrated superior performance when compared to the…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}