PPM/PPM_One: chemical shifts prediction web server

Important: For PPM, prediction on single structure is meaningless. For single structure, please use PPM_One.

The PPM chemical shift predictor was parameterized specifically to account explicitly for motional averaging for its application to molecular ensembles, such as ones obtained from MD simulations. The PPM_one chemical shift predictor was parameterized against PDB structures and includes dynamic averaging effect implicitly

This web server accepts a single PDB file that contains one or multiple protein snapshots with all protons added. It returns predicted backbone (CA, CB, C', HN, N, HA) and proton side chain chemical shift and compares them with experimental chemical shifts (NMR-STAR v2.1 or V3.1 file need to be supplied separately). The program will do sequence alignment between the pdb file and the bmrb file before prediction-experiments comparison.

The web server will write following files:
  1. bmrb_pre.dat: predicted chemical shifts in NMR-STAR v2.1 format
  2. bb_prediction.dat: predicted backbone chemical shift values. File is self-explained.
  3. proton_prediction.dat: predicted proton chemical shift values. File is self explained.
  4. bb_details.dat (proton_details.dat): detailed backbone (and methyl proton) chemical shifts for each snapshot. Each line is for one atom. The first three columns are of AA index, AA name and atom name. The fourth column is the experimental value and the remaining columns are of predicted values of all snapshots. PPM_One doesn't write these two files.

Requirement of input:
  1. Models in PDB file must be separated by ENDMDL keyword, unless there is only one model in the file. For an efficient prediction, the program will ONLY read in residue and atom information from the fist model. That is, all models in the pdb file must have exact same number of atoms and same order of atoms.
  2. All protons must be present in the PDB file. Please see pdb.pdb and bmrb.bmrb as input examples.
  3. Each model can contain multi chains (e.g. for protein protein complex). If so, chain index such as A,B,etc must be given in PDB files. All ligand and water molecules (after TER line in PDB file) will be neglected.
Changing log
  1. Jan. 1st, 2014: initial version
  2. Jul. 10th, 2014: Add support for multiple chain PDB file.
  3. Jan. 10th, 2015: PPM_one is combined into PPM.