mleFit
Fit structures and determine unstructured regions using the maximum likelihood methodology of
D.L. Theobald and D.S. Wuttke, "THESEUS: Maximum likelihood superpositioning and analysis of macromolecular structures," Bioinformatics 22, 2171-2172 (2006).
Usage: mleFit [options] <structures to fit>
where the structures to fit is a space-separated list of pdb filenames, or the name of a MODEL-separated PDB file.
Options are zero or more of:
-selection <atom selection> - which atoms to use in the fit
[default value: name CA]
-mSelect - specify method of regularizing the inverse of
the covariance matrix. By default, a
perturbative approach is used (a small
value is added to the diagonal). If this
option is specified, use diagonal
elements as ``eigenvalues,'' and follow the
scheme of Theoboald and Wuttke.
-verbose - cause intermediate values of log likelihood
and RMSD to be printed.
-pVariance - print a text file (named Variance.txt) with
the variances of all the selected atoms.
-pRTMatrix - print a text file (named RotTrans_Mat.txt)
with the rotation and tranlation matrix for
each structure.
-rmsdThreshold <value> - specify the RMSD value for determining whether
an atom is in an ordered region [default: 1.000000
angstrom]
-writeStructs - the PDB files of the fitted structures are
written to the current working directory
with '.mleFit' appended to the
filename. In addition, another PDB file is
written with the mean coordinates along with
the variance of each atom in the B-factor
column with the file name: MLE-Average.pdb.
Maximum Likelihood RMSD = Sqrt(k/Tr(C^{-1}))
where,
k = number of selected atoms
C = the covariance matrix
Ensemble RMSD = Sqrt(\Sum_{i,j}(delta_{i,j}^2)*1/(3Na*Ns))
Average RMSD = (1/Ns)*\Sum_{i}*
(Sqrt(\Sum_{j}(delta_{i,j}^2))*1/(Na))
where,
delta_{i,j} = |q_{i,j}-qavg_{j}| ,q_{i,j} is the jth atom
in ith structure and qavg_{j} is the jth
atom of the mean structure.
Ns = Number of structures
Na = Number of selected atoms