regarding lmi() calulations

[Kajwal Kumar Patra]

Dear sir, i have tried the lmi approach for calculating cij matrix, here is my code

pdb <- read.pdb("abp.pdb")
dcd <- read.dcd("abp.dcd")
inds <- atom.select(pdb, elety="CA")  #it selects 1948 atoms
trj <- fit.xyz(fixed=pdb$xyz, mobile=dcd, fixed.inds=inds$xyz, mobile.inds=inds$xyz)
cij <- lmi(trj[, inds$xyz])  #i got a 1948*1948 matrix with no negative values 

so the cij matrix contains values in between 0 to 1. where as the same calculation i did with dccm(), i got some -ve values in cij matrix. thats why i am a little confused to follow which calculation to do next analysis like path analysis, community network formations. i think lmi() is more advanced.

1. firstly was my code right for the lmi() calculation? because i saw in the help(lmi) that they defined the object 'trj' as

trj <- fit.xyz(pdb$xyz, dcd, mobile.inds=inds$xyz, fixed.inds=inds$xyz) 

so please make some comments

regards

[Xinqiu Yao]

Hi,

You codes look good! Yes, lmi() returns values between 0 and 1 while dccm() returns between -1 and 1. It is NO PROBLEM to go ahead for network construction, path analysis, community analysis, etc. with either method. Good luck!

[Barry Grant]

not a bug but a Q&A