The data is from the paperDe novo design of peptide binders to conformationally diverse targets with contrastive language modeling"(Published in Science Advances in January 2025) all raw and processed data. The paper published by the Duke University research team constructed a CLIP-based peptide priority screening process PepPrCLIP, which can design short proteins to bind to and destroy previously undruggable pathogenic proteins. Compared with the existing platform RFDiffusion, which uses the target 3D structure to generate peptides, PepPrCLIP is faster and can create peptides that are almost always a better match to the target protein.

Citation

@misc{bhat_2024_13917484, author = {Bhat, Suhaas and Palepu, Kalyan and Hong, Lauren and Mao, Joey and Ye, Tianzheng and Iyer, Rema and Zhao, Lin and Chen, Tianlai and Vincoff, Sophia and Watson, Rio and Wang, Tian Z. and Srijay, Divya and Venkata, Srikar Kavirayuni and Kholina, Kseniia and Goel, Shrey and Vure, Pranay and Desphande, Aniruddha J. and Soderling, Scott H. and DeLisa, Matthew P. and Chatterjee, Pranam}, title = {De Novo Design of Peptide Binders to Conformationally Diverse Targets with Contrastive Language Modeling }, year = 2024, publisher = {Zenodo}, doi = {10.5281/zenodo.13917484}, url = {https://doi.org/10.5281/zenodo.13917484}, }