NanoNet
Rapid and accurate end-to-end nanobody modeling by deep learning
Rapid and accurate end-to-end nanobody modeling by deep learning
NanoNet is a deep learning-based end-to-end tool designed for rapid and accurate modeling of nanobodies (Nbs), monoclonal antibody (mAb) heavy chains, and T-cell receptor (TCR) variable beta (VB) domains. Given an amino acid sequence, NanoNet predicts the 3D coordinates of the backbone and Cβ atoms for the entire variable domain.
NanoNet can be run via our web server, colab notebook, or installed locally. Please take a look at the help section for more details.
Typical run time for 1,000 structures on a single standard CPU:
Backbone + Cβ - less than 5 seconds (For better performance, use GPU)
Backbone + SCWRL - ~20 minutes
Backbone + Modeller - ~80 minutes
Source code: Available at GitHub
Citation: If you use NanoNet please cite: T Cohen, M Halfon, D Schneidman-Duhovny. NanoNet: Rapid and accurate end-to-end nanobody modeling by deep learning. Frontiers in immunology 13, 958584
Contact: tomer dot cohen13 {at} mail dot huji dot ac dot il
Acknowledgement: We are grateful for the support of the Center for Interdisciplinary Data Science Research (CIDR) at The Hebrew University of Jerusalem. In particular, we would like to thank Haimasree Bhattacharya from CIDR for her contributions.