plot.dccm and network analysis

[Aiswarya Pawar]

Hi,

Iam working on one protein with a ligand bound and unbound systems. I have performed MD simulations and

now using Bio3d to solve the correlation between these two systems. I have few doubts regarding the usage of the modules

1. I performed correlation analysis and got an desired plot. But i was wondering if we could increase the plot thickness and make the labels bold and increase thickness. i tried using couple of things lwd , par() function and also box(which="plot",lty="solid") then i figured out scales=list(cex=1.2) but it only scales the labels. Is there any way to change the plot properties.
2. I have performed network analysis, using the following commands

‌

ca.inds <- atom.select(pdb, "calpha")
trj <- read.ncdf("tmp.nc", at.sel=ca.inds,first=300, last=500, stride=2)
cm <-cmap.xyz(trj,dcut=10,pcut=0.75, grpby = pdb$atom[ca.inds$atom,'resno'])
xyz <- fit.xyz(fixed=pdb$xyz,mobile=dcd, fixed.inds=ca.inds$xyz,mobile.inds=ca.inds$xyz)
cij<-dccm(xyz[,ca.inds$xyz],cutoff.cij=0.35)
net <- cna(cij,cm=cm)
plot(net,pdb)

But i encounter few warning messages

Warning messages:
1: In (function (graph, weights = E(graph)$weight, directed = TRUE, :
At community.c:460 :Membership vector will be selected based on the lowest modularity score.
2: In (function (graph, weights = E(graph)$weight, directed = TRUE, :
At community.c:467 :Modularity calculation with weighted edge betweenness community detection might not make sense -- modularity treats edge weights as similarities while edge betwenness treats them as distances

What does these messages mean. How do i overcome this.?

Thanks

[Xinqiu Yao]

Hi,

‌

The correlation plot is based on the lattice package and so regular box() or par() won't work. You may modify some of the properties using the scales and other related arguments (check help(levelplot, lattice) for more detail). Alternatively, if you are not satisfied with some features, export it to a PDF and change directly using Adobe Illustrator.

‌

Both warnings are from latest versions of igraph. The first can be safely ignored. The second won’t generate any troubles as long as you are using a “distance” network. By default, we use the Schulten’s ‘-log(|cij|)’ distance metric, and so there should be no problem if you are following the same line.

[Aiswarya Pawar]

ok. i get it.

I have one more doubt, i tried to visualize the correlation in pymol using pymol.dccm command. I could see the red and blue colored lines drawn between the residues and domains. Do the color relate to correlated and anti correlated? if so which one is for what?

[Xinqiu Yao]

Yes, the color is related with correlation and anti-correlation. In pymol, you can see a few objects on the right that you can turn on or off the display. By those, you can easily tell what are positive or negative.

[Aiswarya Pawar]

Thanks i could figure out.