Creating a `flooding' matrix: mkflood

The program mkflood carries out the principal component analysis on an ensemble of structures (ego-files) and creates the appropriate flooding matrix flooding.lis required to carry out a flooding simulation. Typically, the quality of the flooding matrix will be the better, the larger the used structure ensemble is, i.e., the longer the used MD trajectory is.

To further enhance the accuracy of the flooding matrix, mkflood allows to extrapolate the parameters of the flooding matrix to infinite times. It does so by considering subsets i of the structure ensemble of increasing size $\Delta T_i$ and by performing a principal component analysis on each of these subsets. The function $\lambda(\Delta T)=a-b/\Delta T$ is then fitted to each of the appropriate parameters (the eigenvalues) of the matrices, and an estimate $\lambda(\Delta T\rightarrow\infty)=a$ is obtained. The sub-ensembles used are chosen to increase in size logarithmically, starting with a given minimal size, and up to the total ensemble available.

In `dummy-mode', mkflood allows to generate a file flooding.lis, containing solely a list of selected atoms. This feature is useful to carry out a conventional MD-simulation with rotation/translation correction, as, e.g., required to generate an ensemble for the initial conformational substate.

The format of the program call is

mkflood <pdb-file> <psf-file> <start-num> <end-num> <min. #files> 
        <#steps> <nr_of_dofs> ['<Sel.-string>']

with the following arguments:

<pdb-file>:	Coordinate file used to determine the numbers of the selected atoms.
<psf-file>:	Structure file used to determine the numbers of the selected atoms.
<start-num>:	Number of the first ego-file of the structure ensemble to be used for the principal component analysis.
<end-num>:	Number of the last ego-file used
<min. #files>:	minimal ensemble size used for the extrapolation to infinite times. The first flooding matrix will be computed from the ego-files <start-num>...<start-num>+<min. #files>-1. Attention: In order to avoid singularities (which result in an error message), take care that the minimal ensemble size is larger than the dimension of the configuration space considered, i.e., larger than three times the number of selected atoms.
<#steps>:	Number of flooding matrices to be computed. Each of the matrices is computed from a different number of ego-files, these numbers are distributed logarithmically in the range <min. #files>...<end-num>-<start-num>. Use <#steps>=0 in order to avoid any extrapolation. In this case only one flooding matrix is computed from the complete ensemble given, and the value of <min. #files> is ignored.
<nr_of_dofs>:	The number m of degrees of freedom used as `flooding degrees of freedom' ci. That number must be smaller than or equal to 3N-6, where N is the number of selected atoms.
<Sel.-string>:	Expression to select those atoms from the system, which (a) are used for the translation/rotation correction, (b) which are used for the principal component analysis to create the flooding matrix, and (c) onto which the flooding potential acts upon. A typical selection would include only one representative of groups of atoms, whose motion is strongly correlated (to save computer time), e.g., the alpha carbon of each group, and exclude those atoms of the system, which definitely do not actively take part in the conformational motion to be studied (e.g., solvent atoms).

The general syntax of the selection string is
[+|-]{A|R}{<number|num1-num2|string},
where

+ :	includes the selected atoms/residues (default),
- :	excludes the selected atoms/residues,
A :	refers selection to atom numbers/names,
R :	refers selection to residue numbers/names,
number :	selects a single atom/residue,
num1-num2 :	selects an atom/residue range, and
string :	selects an atom/residue by string-match. Allowed wild-cards are * (string), % (single character), # (string of digits), and + (single digit), similar to the XPLOR selection scheme.

Example: The selection 'AC* AN* -RPRO' selects all carbon (C) atoms and all nitrogen (N) atoms except for those in prolines.

To create a dummy flooding-matrix-file to be used for translation/rotation correction, use:
mkflood <pdb-file> <psf-file> 0 0 0 0 0 ['<Selection-string>']