Paper List

期刊: ArXiv Preprint
发布日期: 2025-11-19
BioinformaticsComputational Biology

Generative design and validation of therapeutic peptides for glioblastoma based on a potential target ATP5A

Shanghai Jiao Tong University | QuietD Biotech

Shikui Tu, Hao Qian, Pu You, Lin Zeng, Jingyuan Zhou, Dengdeng Huang, Kaicheng Li, Lei Xu
Figure
Figure
Figure
Figure
Figure

30秒速读

IN SHORT: This paper addresses the critical bottleneck in therapeutic peptide design: how to efficiently optimize lead peptides with geometric constraints while bridging the gap between computational generation and experimental validation.

核心创新

  • Methodology Introduces POTFlow, the first lead peptide-conditioned flow matching model that incorporates secondary structure priors and optimal transport for shorter, disentangled generation paths
  • Methodology Proposes a dry-to-wet framework that integrates computational design with experimental validation spanning in vitro assays and in vivo PDX models
  • Biology Demonstrates successful optimization of ATP5A-binding peptides for glioblastoma, achieving improved tumor selectivity and in vivo efficacy

主要结论

  • POTFlow outperforms five state-of-the-art methods across multiple metrics, achieving 53.44% similarity, 95.07% compactness, 30.56% affinity, and 1.66Å RMSD on benchmark datasets
  • Generated peptide candidates showed 18-68% higher inhibition of viability rate (IVR) in GBM cells compared to non-cancerous cells (<10%), demonstrating improved tumor selectivity
  • High-dose candidate 4 (20mg/kg) significantly prolonged survival in PDX models (p-value = 0.02) with 40% of mice surviving beyond week 18 compared to 0% in control group
研究空白： Current deep generative methods for peptide design primarily focus on de novo generation from scratch, ignoring the practical need for lead optimization where high sequence similarity correlates with functional similarity, and most lack experimental validation.

摘要: Glioblastoma (GBM) remains the most aggressive tumor, urgently requiring novel therapeutic strategies. Here, we present a dry-to-wet framework combining generative modeling and experimental validation to optimize peptides targeting ATP5A, a potential peptide-binding protein for GBM. Our framework introduces the first lead-conditioned generative model, which focuses exploration on geometrically relevant regions around lead peptides and mitigates the combinatorial complexity of de novo methods. Specifically, we propose POTFlow, a Prior and Optimal Transport-based Flow-matching model for peptide optimization. POTFlow employs secondary structure information (e.g., helix, sheet, loop) as geometric constraints, which are further refined by optimal transport to produce shorter flow paths. With this design, our method achieves state-of-the-art performance compared with five popular approaches. When applied to GBM, our method generates peptides that selectively inhibit cell viability and significantly prolong survival in a patient-derived xenograft (PDX) model. As the first lead peptide-conditioned flow matching model, POTFlow holds strong potential as a generalizable framework for therapeutic peptide design.