Why does RMSIP value show greater than 1?

[Yazhini arangas]

I am running all aaenma and using gap.inspect option to compare modes of proteins with varying length. When I calculate RMSIP value, why am I getting value greater than 1. What could be the reason? Is it because of vector length mismatch.

[Xinqiu Yao]

Hi,

Can you tell what exactly the commands you used?

[Yazhini arangas]

I am using the following code.

library(bio3d)
a <- pdbaln(c("4YUG.pdb","4YUG_s1_bp.pdb","4YUG_1_tf.pdb"))
b <- read.all(a)
gaps.pos <- gap.inspect(b$all)
modes <- aanma.pdbs(pdbs = b, outmodes='noh')
modes1 <- matrix(NA, nrow=ncol(b$all), ncol=ncol(modes[[1]]$U))
modes2 <- matrix(NA, nrow=ncol(b$all), ncol=ncol(modes[[2]]$U))
modes3 <- matrix(NA, nrow=ncol(b$all), ncol=ncol(modes[[3]]$U))
modes1[!is.gap(b$all[1, ]), ] <- modes[[1]]$U
modes2[!is.gap(b$all[2, ]), ] <- modes[[2]]$U
modes3[!is.gap(b$all[3, ]), ] <- modes[[3]]$U
rm1 <- rmsip(modes1[gaps.pos$f.inds, ], modes2[gaps.pos$f.inds, ], subset=10)
rm2 <- rmsip(modes1[gaps.pos$f.inds, ], modes3[gaps.pos$f.inds, ], subset=10)
print(rm2$rmsip)

I have used gaps.pos$f.inds to avoid dimension mismatch error in rmsip calculation for 4YUG and 4YUG_s1_bp protein pair. Whereas there is no gaps in the alignment between 4YUG and 4YUG_1_tf pairs. When I calculate rmsip for 4YUG and 4YUG_1_tf pair, it gives value greater than 1. Since I am new to this work, it is totally perplexing.

[Lars Skjærven]

The modes are no longer orthogonal when subsetting like this. You need to calculate the effective hessian and diagonalize it. This is done internally in aanma.pdbs() (and nma.pdbs()) when outmodes = "calpha", but not when outmodes = "noh" as aligning heavy atoms is not implemented.

[Yazhini arangas]

Thank you very much for the clarification. I can now understand my mistake.

[Xinqiu Yao]

Thanks for explaining, Lars. What about normalized the modes? For example normalize.vector(modes1[gaps.pos$f.inds, ]), and then using normalized modes to do rmsip() should give a value below 1.