Data & Visualizations
Visualizations for the output job data.
Rank | MPNN Score | RMSD | Mean pLDDT | Max PAE | pTM |
---|
Our RFdiffusion implementation predicts protein structures and runs them through ProteinMPNN to generate 200 sequences. The best 5 ProteinMPNN sequences (sorted by MPNN Scores), are then passed through AlphaFold2.
Important Metrics
Higher values are correlated with greater accuracy.
Lower ProteinMPNN scores are correlated with greater accuracy.
2D animation of the protein diffusion process over each timestep.
Predicted Aligned Error (PAE)
ProteinMPNN Predicted Sequences
RFdiffusion only predicts the backbone of the protein. Inverse folding models like ProteinMPNN are required to predict sequences that can best fold into the RFdiffusion predicted backbone.
NOTE: Lower ProteinMPNN scores are better.
AI models produce responses and outputs through sophisticated algorithms and learning techniques, which may result in inaccuracies. By engaging with this model, you accept responsibility for any potential harm resulting from its responses or outputs.
Config
The configuration that was used for this job.
Configuration Setting | Set Value |
---|---|
Binder / Scaffolding Structure | none |
Contig Input | 240-240 |
Hotspots (optional) | none |
Symmetry | tetrahedral |
Timesteps | 100 |
Files
The following files were either used as input(s) or produced by this job.
Output Files
Download all as a zip file:
Citations
Please cite the following if you wish to publish data produced from this job.
Neurosnap Inc. - Computational Biology Platform for Research. Wilmington, DE, 2022. https://neurosnap.ai/.