dccm cross correlation length mismatch

[Mamta Mohan]

Hi,

I am following tutorial for 'Trajectory Analysis with Bio3D'.

Instead of example pdb. I am using my pdb which has two residues label with MTSL spin label.

Initially pdb read gives warning which I wrote a while back but It was advised to ignore them since being warning.

Since I am working (as per tutorial only with carbon-backbone). No further errors or warnings were generated.

But for Cross-correlation plot I am getting following error:

3D visualization of Cross-Correlations

pymol.dccm(cij, pdb, type="launch") Error in pymol.dccm(cij, pdb, type = "launch") : unequal vector lengths sum(pdb$calpha) [1] 162

I checked pdb it is 162 and cij is dccm [1:164, 1:164]

So mismatch is clear.

Is there a way to solve this.

[Xinqiu Yao]

Hi,

Can you provide an example pdb file to reproduce your errors?

[Mamta Mohan]

Please find attached pdb and psf files.

ionized.pdb and ionized.psf are solvated system pdb and psf file.

stripped.pdb and stripped.psf are files with out solvent.

[Lars Skjærven]

Hi Mamta, Make sure you are doing the correct atom selections here. Looks like you have non-standard residues in your protein (CYR1) which atom.select(pdb, "calpha") does not recognize.

> pdb <- read.pdb("stripped.pdb")
> pdb
 Call:  read.pdb(file = "stripped.pdb")

   Total Models#: 1
     Total Atoms#: 1339,  XYZs#: 4017  Chains#: 2  (values: P I)

     Protein Atoms#: 1294  (residues/Calpha atoms#: 162)
     Nucleic acid Atoms#: 0  (residues/phosphate atoms#: 0)

     Non-protein/nucleic Atoms#: 45  (residues: 11)
     Non-protein/nucleic resid values: [ CLA (9), CYR1 (2) ]

Calling atom.select(pdb, "calpha") you select only those CA atoms with known residue names. In this case, you can use the elety argument:

# select only CA atoms
> sele <- atom.select(pdb, elety="CA")
> sele
 Call:  atom.select.pdb(pdb = pdb, elety = "CA")

   Atom Indices#: 164  ($atom)
   XYZ  Indices#: 492  ($xyz)

+ attr: atom, xyz, cal

# calculate normal modes
> m <- nma(pdb, inds=sele, mass=FALSE)
 Building Hessian...            Done in 0.061 seconds.
 Diagonalizing Hessian...       Done in 0.094 seconds.

# check dimensions of resulting modes object
> dim(m$U)
[1] 492 492

# calculate cross correlations
> cij <- dccm(m)
  |======================================================================| 100%

# dimensions.... 
> dim(cij)
[1] 164 164

Here I used NMA as an example, but the same goes for reading a trajectory: trj[, sele$xyz].

Good luck. Lars

[Lars Skjærven]

Side note: The following error message is not very useful in this case:

> m <- nma(pdb, inds=sele, mass=T)
Error in aa2mass(pdb = c("MET", "ASN", "ILE", "PHE", "GLU", "MET", "LEU",  : 
  must supply 'pdb' object or vector of amino acid residue names

> aa2mass(pdb$atom$resid[sele$atom])
Error in aa2mass(pdb$atom$resid[sele$atom]) : 
  must supply 'pdb' object or vector of amino acid residue names

> any(nchar(pdb$atom$resid[sele$atom]) > 3)
[1] TRUE

should check line 19 of aa2mass.R.

[Mamta Mohan]

Dear Lars,

Thank you for your response.

I understand your explanation about backbone selection.

The reason I am doing backbone selection is because I do not know how to incorporate my labels generated by function pdb when input file is provided.

If I can do that I am really interested to analyse the labels and surrounding neighbours.

Is there a way to incorporate labels?

If

[Lars Skjærven]

We normally use c-alpha atoms for cross correlation analysis. Using the atom selection statements above should be fine. Just make sure you're matrix is the correct dimensions after calling dccm().

[Mamta Mohan]

Thank you for your response.

I will do as suggested.

What I see is C-alpha carbon is not taken into account when I try to load pdb file via read.pdb function

Protein has 164 C-alpha atoms, at two places where I modified amnio-acid side-chain to attach label those two C-alpha are not accounted for. They are skipped. And only 162 C-alpha atoms are read and parsed.

So analysis as I do the way you suggested will not reveal the real NM and Cross-Correlations because it is done on 162 rather than 164.

So if I can address this during pdb read function, I think even analysis on C-alpha backbone will be very helpful to me.

Is there a way I can modify this call.

I will appreciate your response.

[Lars Skjærven]

Hi Mamta, Sorry but that's not correct. read.pdbreads ALL atoms in the pdb file. It is your atom selection which is wrong here. As I tried to demonstrate above, the automatic calpha selection keyword will not work here because of the unknown residue name of your spin label. Using elety="CA" will however successfully select all 164 calpha atoms. Also as stated above, the corresponding modes object contains 3x164 modes- indicating that all 164 calpha atoms are used in the modes calculation.

[Mamta Mohan]

Dear Lars,

Thank you for your explanation. I understand now how selection is working.

I will use elety= "CA" to build model.