Non-linear torsional pair correlation

[Former user]

First off, let me start by saying that I love this package. The linear dynamic cross-correlation analysis has helped me reveal many hidden features on my protein systems.

I wonder if there is any possibility that non-linear torsion pair correlations have been implemented for side chains or will be implemented in the package by any chance. I came across the function torsion.mi recently: https://github.com/Grantlab/bio3d/blob/master/new_funs/torsion.mi.R. Could you please let me know how to include these functions into the bio3d package, including xyz2torsion, big.matrix, mclapply etc.?

[Xinqiu Yao]

Hi,

Thanks for your comments on Bio3D.

Those functions under 'new_funs/' are experimental. Although in most cases they should work fine, but I am not 100% sure because they might not be updated promptly. Another issue about these experimental functions is that they are usually not well documented.

We will gradually move some "new_funs" into the main package but an exact timeline is unsure. We always encourage users to make their own contributions to Bio3D. So, I recommend play around with the functions, modify whatever you think necessary, and may send your version back to us with detailed examples to run it for further testing. You could attach your files here or send them by Email to [e-mail].

Let me know if you have any further questions.

[Barry Grant]

This would indeed be useful and I think we should migrate this torsion based analysis function (along with a related visualization feature) into the main package.

Would be useful to hear what you are after with your envisaged side-chain analysis (and how close this function gets) so we can perhaps focus developments in this area.

[Shayon Bhattacharya]

Hi Xin and Barry,

Thank you for your interest and valuable insights for the feature in question. To answer Barry, I am looking for whether 'certain' (polar/non-polar) side chains have highly correlated motions, which might be of interest when studying specific folding (or for that matter unfolding) patterns.

[Shayon Bhattacharya]

Any update on this please?

[Xinqiu Yao]

Hi,

We are still working on integration.

Regarding your original question, simply source the files and load proper libraries should work. For example,

library(bio3d)
library(parallel)
library(bigmemory)

source('torsion.mi.R')
source('xyz2torsion.R')

pdb <- read.pdb('test.pdb')
xyz <- read.ncdf('test.nc')

cij <- torsion.mi(xyz, pdb)

Let me know if you have any more questions.

[Barry Grant]

Is it time for a CRAN patch release? What do you think Xinqiu?

[Xinqiu Yao]

Yes, I think it is time. We have already several bugfixes since last release, some related to core functions.

Will go over remaining open bug-related issues shortly and see if I can fix them quickly.

[Xinqiu Yao]

Just pick up this again. We have more bugfixes/minor updates on the releases branch now. The only thing left is to merge the bugfix_readpdb branch (See description of the problem here). I hope you can help test the branch and, if there is no problem, we merge the branch and release a new patch. What do you think?

[Barry Grant]

We must release a new version soon. The read.pdb fix you note together with the hmmer and blast function fixes should be pushed out ASAP!