Prot-LAMBDA

We provide the different predictions made by our Prot-LAMBDA model as presented in our paper Prot-LAMBDA: Explicit Distance Learning Enhances Structural Reasoning in Protein Language Models predicted_contact_maps.p : Predicted contact maps for 50 CASP14 proteins and 194 CAMEO (Between April 1 to June 25, 2022) proteins. predicted_distograms.p : Predicted 64 bin distograms for 30 non-redundant proteins from CASP14+CAMEO, and proteins corresponding to 37 FM and FM/TBM domains from CASP14. predicted_structures.zip : Predicted 3D structures in PDB format for 50 CASP14 proteins and 194 CAMEO (Between April 1 to June 25, 2022) proteins. Contains the intermediate predictions of stages. predicted_structures_rag.zip : Predicted 3D structures in PDB format for 90 proteins from CASP14+CAMEO, having sufficient template coverage. The trained model weights can be found in https://huggingface.co/KiharaLab/ProtLAMBDA The source codes are available in https://github.com/kiharalab/Prot-LAMBDA/ Citation: Ibtehaz, N., Zhang, Z., Kagaya, Y., Xu, M., Tomii, K., & Kihara, D. Prot-LAMBDA : Explicit Distance Learning Enhances Structural Reasoning in Protein Language Models. (In submission)