HD-AIP: A Heterogeneous Dual-Stream Alignment-Free Framework for Anti-Inflammatory Peptide Prediction Based on Language Models and CT-Net.

HD-AIP achieves superior anti-inflammatory peptide prediction accuracy by integrating dual-stream, alignment-free features from language models and CT-Net, outperforming baseline models.

• Why it matters: Accurate computational screening of anti-inflammatory peptides is crucial for developing therapies for chronic and autoimmune diseases, but traditional methods struggle with short sequences and high computational costs.
• What they did: The approach employs a heterogeneous dual-stream architecture extracting features from macro- and microscopic perspectives using large protein language models, physicochemical properties, BioVec embeddings, CNN, and Transformer networks, combined with a dynamic soft ensemble.
• The result: HD-AIP surpasses existing models on multiple metrics, enabling efficient high-throughput virtual screening and candidate discovery of anti-inflammatory peptides without relying on sequence alignment.