Creating Protein Structure Networks

[Yu Zhu]

Hello, I’ve read the tutorials about Protein Structure Network and I tried to use my own data but failed.

Before I use the bio3D, I’ve finished Correlation Analysis and got a matrix in txt format which means I want to skip some steps in the tutorial to reach my goal. PDB file can be correctly read while my matrix can’t. I think some changes in the tutorials need to be done in the following lines:

modes <- nma(pdb)
cij <- dccm(modes)
net <- cna(cij, cutoff.cij=0.3)

However, I’m still a green hand in R and cannot correctly shift my matrix.txt file into the cij variable. I wish someone could help me solve the issue.

Thx a lot.

[Xinqiu Yao]

Hi,

‌

Try cij <- as.matrix(read.table("cij.txt"))

[Yu Zhu]

Yes, I’ve tried this but something is still wrong. I couldn’t get the net plot as the picture shown below.

I checked the class(cij) in the tutorials example and it turned out to be “dccm” “matrix” while mine result is just “matrix”. I don’t know whether this is the key to the problem.

Attached with the comment is part of my matrix since I wonder whether something is wrong with my matrix file and hope you can help me find the solution.

Thank you.

[Yu Zhu]

part of my matrix file

[Xinqiu Yao]

The class of matrix should be fine.

Try just typing net, what does it tell you? If no error/warning, the network object should be fine. And the problem must be from pdb.

Try just plot(net). Do you see a plot?

One possibility is the pdb contain non-protein atoms, or the number of residues does not match the number of nodes in the network.

‌

Let me know if it helps.

[Yu Zhu]

Yes, it worked after I removed some ions in the pdb file. Besides, I wonder whether some non-standard amino acids in the pdb file could affect the accuracy of the result.

Hope to hear from you soon.

Thank you.

[Xinqiu Yao]

Bio3D can recognize non-standard amino acids.

Type pdb and see if it says Non-protein/nucleic atoms#: 0.

If non-standard amino acids are expected to be in the network, the result should be fine.

‌

‌

[Yu Zhu]

By typing pdb, I got this:

Non-protein/nucleic Atoms#: 221 (residues: 20)
Non-protein/nucleic resid values: [ CY1 (8), CY2 (3), CY3 (6), HD1 (2), HE1 (1) ]

These are some residues closely similar to CYS and HIS, differing only a few hydrogen ions. Is that necessary to manually edit the pdb file or could you please give me a better solution.

Thanks.

[Xinqiu Yao]

If you didn’t get any error/warning, I don’t think you need to change the names. The function distinguishes residues by residue numbers and extracts C-alpha coordinates to build the layout of the plot. I assume all those unrecognized residues have C-alpha, right?

‌

You may change the residue name for double checking anyway. For example,

pdb$atom$resid[pdb$atom$resid=="CY1"] <- "CYS"
...

‌

[Yu Zhu]

I think it’s working properly now.

Thank you very much.