List of Tables

• The clusters, hinges, and related dihedrals of Met-enkephalin, shown in .
• Proteins and peptides used in NEIMO simulations. The structures listed are the initial Protein Database files, except for the peptides ``MEnk'' and ``Ala9,'' which were created using the BIOGRAF peptide builder.
• Timing results for the ten protein/peptides systems studied here. The average times per timestep of the NEIMO calculation and the nonbond calculation are given, along with the NEIMO time divided by , and the nonbond time divided by the number of atoms, .
• The average values of the Met-enkephalin dihedrals from 5 ps NEIMO () and Cartesian () dynamics simulations, compared to the initial values and compared to each other.
• Crystal structures used in the H64 dataset.
• The number of samples of the three residue types in the protein datasets.
• This table lists the number of (and percentage of the maximum possible) gridpoints which have non-zero values for each of the three residue types at different grid spacings, .
• This table lists the number of high-probability gridpoints: the number which have probabilities above .
• This highest-probability gridpoint of each residue type for each grid spacing.
• The percentage of sample 's falling within each quadrant of conformational space.
• The distribution of secondary structure designations in the crystal structures of the SS58 dataset.
• The number of PGMC dihedrals () for each amino acid, the number of occurrences of each amino in the H64 crystal structures, and the number of populated (non-zero) gridpoints at each grid spacing. for alanine and glycine. The numbers in italics are more than 95%of the sample size, indicating that nearly every conformation occupies a different gridpoint.
• The total number of gridpoints for dihedrals at different grid spacings.
• The number of populated gridpoints in the , distribution of amino acids with . Lysine appears to be the most flexible while tryptophan is the least flexible.
• The highest-probability gridpoint for each amino acid for . is the probability of this particular conformation.
• Conformations generated for Met-enkephalin from peak , , and gridpoints at different grid spacings. Because of steric overlap, the energy of the 5 conformation is greater than kcal/mol.
• Conformations generated from peak , , and gridpoints at different grid spacings, after minimization with the DREIDING forcefield. The energy of each conformation is given, along with the RMS deviation from its original (un-minimized) structure.
• LS is the structure produced by rotating the dihedrals of our starting structure to the values reported for the ECEPP/2 global minimum [72]. LS was obtained by a Cartesian-coordinate conjugate-gradients minimization of LS using the DREIDING forcefield.
• Dihedral numbering used in and .
• regions indicating residue is likely to be in an helix or sheet conformation; i.e., its and fall within the corresponding region listed in the lower table. and are defined in .
• Variables used in PGMC C Builder. For Phase 1, ``steps'' refers to the number of conformations sampled as each residue is added. For Phase 2, it refers to the total number of conformations sampled.
• Values used for production runs of the C Builder.
• The energy, rms deviation in backbone atoms (RMSB), and rms deviation in dihedrals (RMSD) for each of the 20 backbone conformations generated for crambin.
• The energy and rms deviation in atomic coordinates for each of the crambin models produced by the PGMC C Builder.
• RMS deviations for different regions of the crambin model.
• The rms deviations in various types of dihedrals for the crambin model and the percentage of each type of dihedral with deviations less than 30 or more than 90.
• The proteins modeled by the PGMC C Builder. The reference crystal structure is given along with the number of residues in the protein and the percent of these which are in helices and sheets.
• The results from Phase 1 constructions of the backbone conformations of several proteins.
• The results from Phase 2 constructions of the sidechains of crambin, plastocyanin, and flavodoxin.
• A comparison of the results for flavodoxin vs. other methods. ``Correct'' refers to dihedrals predicted to within 20 of their crystal structure values.
• Equilibrium geometries and force constants in the CFF.
• Best-fit fcc lattice conformations of several proteins, before and after minimization with the CFF.
• Results from building all-atom conformations from various C conformations of crambin using the PGMC C Builder.
• The loop residues of HyHEL-5 and McPC603.
• The rms deviations for all atoms, backbone atom (BB), and C coordinates are given for the predicted conformations of the McPC603 hypervariable loops.
• The lowest-energy loops created in phases 1 and 2 of simulations of HyHEL-5.
• The best rms deviation of C coordinates from the crystal structure from among the 1000 loops generated in Phase 1.
• The results from this work (PGMC) compared to results from three different methods: a conformational-search algorithm[118], a method which uses the canonical structures of loops other than H3 (L3 of HyHEL-5 also does not fit one of the canonical structures)[116], and a method which combines conformational searching with comparisons to database conformations[108].