Unusual RMSD values from struct.aln

[Former user]

Hi ,

I have to compare the structure of my favorite protein to the structures of related proteins. I am trying to use struct.aln to achieve this. However, in some cases, the final rmsd value is too high. For example,

p <- read.pdb("1E2D.pdb") q <- read.pdb("3TMK.pdb") z <- struct.aln(p,q,max.cycles = 100, cutoff = 0.5) rmsd_final <- tail (z$rmsd,1)

The rmsd_final value is get for this particular case is 17.511. Using the align function in pymol, I get a value of 0.696.

Any suggestions ?

Thanks, K.

[Lars Skjærven]

I'm not sure this is a bug. The initial superposition in struct.aln() is based on sequence alignment using muscle. In your case 3TMK includes 8 chains, and it looks like the sequence alignment is messed up because of this. See the output from:

pdbaln(c("1E2D", "3TMK"))

Consider trimming to relevant chains prior to calling struct.aln(), or, if relevant, align all chains in pdb:

> files <- get.pdb(c("1E2D", "3TMK"), split=TRUE)
> pdbs <- pdbaln(files, fit=TRUE)
> rd <- rmsd(pdbs, fit=TRUE)
> rd
       [,1]  [,2]  [,3]  [,4]  [,5]  [,6]  [,7]  [,8]  [,9]
 [1,] 0.000 1.781 1.710 2.052 1.723 1.797 1.718 1.785 1.713
 [2,] 1.781 0.000 0.873 1.425 0.502 0.204 0.497 0.168 0.492
 [3,] 1.710 0.873 0.000 1.396 0.722 0.931 0.754 0.899 0.717
 [4,] 2.052 1.425 1.396 0.000 1.187 1.477 1.156 1.475 1.188
 [5,] 1.723 0.502 0.722 1.187 0.000 0.563 0.146 0.542 0.110
 [6,] 1.797 0.204 0.931 1.477 0.563 0.000 0.558 0.111 0.564
 [7,] 1.718 0.497 0.754 1.156 0.146 0.558 0.000 0.539 0.157
 [8,] 1.785 0.168 0.899 1.475 0.542 0.111 0.539 0.000 0.540
 [9,] 1.713 0.492 0.717 1.188 0.110 0.564 0.157 0.540 0.000

[Lars Skjærven]

reopen if needed