`plot.matrix.loadings.Rd`

Plot residue-residue matrix loadings of a particular PC that is obtained from a principal component analysis (PCA) of cross-correlation or distance matrices.

# S3 method for matrix.loadings plot(x, pc = 1, resno = NULL, sse = NULL, mask.n = 0, plot = TRUE, ...)

x | the results of PCA as obtained from |
---|---|

pc | the principal component along which the loadings will be shown. |

resno | numerical vector or ‘pdb’ object as obtained from |

sse | a ‘sse’ object as obtained from |

mask.n | the number of elements from the diagonal to be masked from output. |

plot | logical, if FALSE no plot will be shown. |

... | additional arguments passed to |

Plot and also returns a numeric matrix containing the loadings.

The function plots loadings (the eigenvectors) of PCA performed on a set of matrices
such as distance matrices from an ensemble of crystallographic structures
and residue-residue cross-correlations or covariance matrices derived from
ensemble NMA or MD simulation replicates (See `pca.array`

for detail).
Loadings are displayed as a matrix with dimension the same as the input matrices
of the PCA. Each element of loadings represents the proportion that the corresponding
residue pair contributes to the variance in a particular PC. The plot can be used
to identify key regions that best explain the variance of underlying matrices.

Xin-Qiu Yao

Skjaerven, L. et al. (2014) *BMC Bioinformatics* **15**, 399.
Grant, B.J. et al. (2006) *Bioinformatics* **22**, 2695--2696.

if (FALSE) { attach(transducin) gaps.res <- gap.inspect(pdbs$ali) sse <- pdbs$sse[1, gaps.res$f.inds] # calculate modes modes <- nma(pdbs, ncore=NULL) # calculate cross-correlation matrices from the modes cijs <- dccm(modes, ncore=NULL)$all.dccm # do PCA on cross-correlation matrices pc <- pca.array(cijs) # plot loadings l <- plot.matrix.loadings(pc, sse=sse) l[1:10, 1:10] # plot loadings with elements 10-residue separated from diagonal masked plot.matrix.loadings(pc, sse=sse, mask.n=10) }