Method detecting surface / bounday Carbon-alpha (or sidechains) atoms of PDB 3d coordinates / points

Question

[andres.susrud]

Hi,

I'm looking for a method or preferrably an R package that can help me with detecting the amino acids (carbon alpha or sidechains) that are surface exposed in a pdf 3d coordinate file.

appreciate any feedback on methods to implement or any reference to R packages with these possibilities built in.

quick searches brings me in the direction of point clound boundry detection.

thanks,

Andres

Answer 1

[Mensur Dlakic]

IMO, one of the most useful programs in protein structure analysis is DSSP. It can do what you want, but you will need to post-process its output.

Here is a partial DSSP output from this PDB structure:

  #  RESIDUE AA STRUCTURE BP1 BP2  ACC     N-H-->O    O-->H-N    N-H-->O    O-->H-N    TCO  KAPPA ALPHA  PHI   PSI    X-CA   Y-CA   Z-CA
    1    1 A M              0   0   67      0, 0.0   265,-2.4     0, 0.0     2,-0.4   0.000 360.0 360.0 360.0 117.0   21.2   23.0   18.3
    2    2 A K  E     -A  265   0A  35    263,-0.2    27,-1.8    24,-0.2    26,-1.5  -0.889 360.0-164.9-104.4 134.9   23.1   20.0   16.9
    3    3 A F  E     -Ab 264  29A   1    261,-2.7   261,-2.0    -2,-0.4     2,-0.4  -0.934   3.6-170.9-118.5 137.9   25.0   20.0   13.7
    4    4 A V  E     -Ab 263  30A   1     25,-2.1    27,-2.8    -2,-0.4     2,-0.4  -0.988   3.5-165.0-129.0 134.2   26.3   17.1   11.7
    5    5 A S  E     +Ab 262  31A   0    257,-2.9   257,-2.5    -2,-0.4     2,-0.3  -0.977  14.2 173.5-116.9 135.8   28.7   17.3    8.7
    6    6 A F  E     - b   0  32A   0     25,-2.6    27,-1.8    -2,-0.4     2,-1.1  -0.944  31.5-151.3-151.1 120.8   29.1   14.2    6.5
    7    7 A N  E     - b   0  33A   0     -2,-0.3   252,-0.5    25,-0.2    27,-0.2  -0.838  23.5-177.3 -87.8 101.4   30.9   13.7    3.2
    8    8 A I        -     0   0    0     25,-1.4    -1,-0.2    -2,-1.1    26,-0.2   0.648  21.2-146.1 -78.3 -16.9   28.6   10.9    2.0
    9    9 A N  S    S+     0   0   16     24,-1.0    26,-0.2   248,-0.1    25,-0.1   0.952  82.6  12.6  47.1  62.2   30.5   10.3   -1.2
   10   10 A G    >>  -     0   0   12     24,-2.7     4,-1.0    23,-0.2     3,-0.6   0.766  69.0-159.2 105.1  82.3   27.5    9.4   -3.2
   11   11 A L  G >4 S+     0   0    5     24,-1.7     3,-0.6     1,-0.2     7,-0.2   0.873  92.3  49.0 -57.5 -43.1   24.1   10.2   -1.6
   12   12 A R  G 34 S+     0   0  138     23,-0.6    -1,-0.2     1,-0.2    24,-0.1   0.762 108.8  54.5 -73.0 -19.6   22.0    7.8   -3.7

The column labeled ACC (9th from the left) contains absolute solvent accessibility in squared angstroms (columns 36-38 in DSSP output). If you divide those numbers by maximum solvent accessibility of each amino acids (those numbers are available in this paper), you will get relative solvent accessibility for each residue. Excluding the first two residues listed above, these are phi-psi angles and relative accessibilities for the next 10 residues:

231.5   237.9   0.0051  E
221.0   234.2   0.0070  E
233.1   235.8   0.0000  E
198.9   220.8   0.0000  E
262.2   201.4   0.0000  E
271.7    83.1   0.0000  C
 37.1   162.2   0.1019  C
 95.1   182.3   0.1429  C
292.5    56.9   0.0305  H
277.0    80.4   0.5565  H

As you can see, fully paired strand residues tend to have near-zero solvent accessibility. I extract these numbers using a custom program which is not portable, but BioPython has a module that can process DSSP output files.