EVfold Predict Unknown 3D Structure for individual protein domains


Please Note:    if your protein has a known 3D structure or is in a family which has a known homologous structure in the PDB, try using EVcouplings or Protein Model Portal. Also, if your protein of interest is multi-domain, please select a single domain for analysis.

Basic Settings

 ( Helpful resource: proteinmodelportal.org )

Advanced Settings
We provide 3 options for sequence alignment, please choose ONE from below.
Please Note: The quality of your results depends on the quality and depth of the alignment. If initial results do not make sense, please check the e-value table to ensure you have good coverage and, in the case of generated alignments, ensure you have used a suitable e-value threshold.
This option lets you upload a (FASTA formatted) alignment file. Among the sequences in the alignment, exactly one must match the protein identifier entered above. If you pasted a protein sequence above, the Job Name will be used for the protein identifier. See the help page for more details.
Please Note: make sure file size < 100M
Enter the PFAM accession number here and the server will retrieve the multiple sequence alignment from the PFAM database. Also enter a Uniprot name to select the correct family member from the alignment.
Proteins for publications were run with 95% cutoff for conservation and 10/i weights for ranking ECs. Weights are only implemented for simulated annealing stage in CNS not in the distance geometry calculations and have only minor effects on our protocol.
Please Note: Overriding these limits may lead to very long job completion times and/or low likelihood of good quality results.