Distance Matrix Analysis

Construct a distance matrix for a given protein structure.

dm(...)

# S3 method for pdb
dm(pdb, inds = NULL, grp = TRUE, verbose=TRUE, ...)
# S3 method for pdbs
dm(pdbs, rm.gaps=FALSE, all.atom=FALSE, 
                        aligned.atoms.only=NULL, ...)

# S3 method for xyz
dm(xyz, grpby = NULL, scut = NULL, mask.lower = TRUE, 
                  gc.first=FALSE, ncore=1, ...)

Arguments

pdb	a pdb structure object as returned by read.pdb or a numeric vector of ‘xyz’ coordinates.
inds	atom and xyz coordinate indices obtained from atom.select that selects the elements of pdb upon which the calculation should be based.
grp	logical, if TRUE atomic distances will be grouped according to their residue membership. See ‘grpby’.
verbose	logical, if TRUE possible warnings are printed.
pdbs	a ‘pdbs’ object as returned by read.fasta.pdb, read.all, or pdbaln.
rm.gaps	logical, if TRUE gapped positions are removed in the returned value.
all.atom	logical, if TRUE all-atom coordinates from read.all are used.
aligned.atoms.only	logical, if TRUE only equivalent (aligned) atoms are considered. Only meaningful when all.atom=TRUE. Default: FALSE.
xyz	a numeric vector or matrix of Cartesian coordinates.
grpby	a vector counting connective duplicated elements that indicate the elements of xyz that should be considered as a group (e.g. atoms from a particular residue).
scut	a cutoff neighbour value which has the effect of excluding atoms, or groups, that are sequentially within this value.
mask.lower	logical, if TRUE the lower matrix elements (i.e. those below the diagonal) are returned as NA.
gc.first	logical, if TRUE will call gc() first before calculation of distance matrix. This is to solve the memory overload problem when ncore > 1 and xyz has many rows/columns, with a bit sacrifice on speed.
ncore	number of CPU cores used to do the calculation. ncore>1 requires package ‘parallel’ installed.
...	arguments passed to and from functions.

Details

Distance matrices, also called distance plots or distance maps, are an established means of describing and comparing protein conformations (e.g. Phillips, 1970; Holm, 1993).

A distance matrix is a 2D representation of 3D structure that is independent of the coordinate reference frame and, ignoring chirality, contains enough information to reconstruct the 3D Cartesian coordinates (e.g. Havel, 1983).

Value

Returns a numeric matrix of class "dmat", with all N by N distances, where N is the number of selected atoms. With multiple frames the output is provided in a three dimensional array.

References

Grant, B.J. et al. (2006) Bioinformatics 22, 2695--2696.

Phillips (1970) Biochem. Soc. Symp. 31, 11--28.

Holm (1993) J. Mol. Biol. 233, 123--138.

Havel (1983) Bull. Math. Biol. 45, 665--720.

Author

Barry Grant

Note

The input selection can be any character string or pattern interpretable by the function atom.select. For example, shortcuts "calpha", "back", "all" and selection strings of the form /segment/chain/residue number/residue name/element number/element name/; see atom.select for details.

If a coordinate vector is provided as input (rather than a pdb object) the selection option is redundant and the input vector should be pruned instead to include only desired positions.

See also

plot.dmat, read.pdb, atom.select

Examples

# \donttest{
# PDB server connection required - testing excluded

##--- Distance Matrix Plot
pdb <- read.pdb( "4q21" )
#>   Note: Accessing on-line PDB file
#> Warning: /var/folders/xf/qznxnpf91vb1wm4xwgnbt0xr0000gn/T//Rtmp4WslmZ/4q21.pdb  exists. Skipping download
k <- dm(pdb,inds="calpha")
filled.contour(k, nlevels = 10)

## NOTE: FOLLOWING EXAMPLE NEEDS MUSCLE INSTALLED
if(check.utility("muscle")) {

##--- DDM: Difference Distance Matrix
# Downlaod and align two PDB files
pdbs <- pdbaln( get.pdb( c( "4q21", "521p"), path = tempdir() ), outfile = tempfile() )

# Get distance matrix
a <- dm.xyz(pdbs$xyz[1,])
b <- dm.xyz(pdbs$xyz[2,])

# Calculate DDM
c <- a - b

# Plot DDM
plot(c,key=FALSE, grid=FALSE)

plot(c, axis.tick.space=10,
     resnum.1=pdbs$resno[1,],
     resnum.2=pdbs$resno[2,],
     grid.col="black",
     xlab="Residue No. (4q21)", ylab="Residue No. (521p)")
}
#> Warning: /var/folders/xf/qznxnpf91vb1wm4xwgnbt0xr0000gn/T//Rtmp4WslmZ/4q21.pdb  exists. Skipping download
#> Reading PDB files:
#> /var/folders/xf/qznxnpf91vb1wm4xwgnbt0xr0000gn/T//Rtmp4WslmZ/4q21.pdb
#> /var/folders/xf/qznxnpf91vb1wm4xwgnbt0xr0000gn/T//Rtmp4WslmZ/521p.pdb
#> ..
#> 
#> Extracting sequences
#> 
#> pdb/seq: 1   name: /var/folders/xf/qznxnpf91vb1wm4xwgnbt0xr0000gn/T//Rtmp4WslmZ/4q21.pdb 
#> pdb/seq: 2   name: /var/folders/xf/qznxnpf91vb1wm4xwgnbt0xr0000gn/T//Rtmp4WslmZ/521p.pdb 
# }

if (FALSE) {
##-- Residue-wise distance matrix based on the
##   minimal distance between all available atoms
l <- dm.xyz(pdb$xyz, grpby=pdb$atom[,"resno"], scut=3)
}