[Chimera-users] Contact Maps (RRdistance Maps) questions:

Elaine Meng meng at cgl.ucsf.edu
Wed Mar 22 19:05:58 PDT 2017


Hi Nicky,
Unfortunately the RR Dist Maps tool is limited in just the ways you describe, so not that useful for what you want to do.  A better way is with “findclash”  but that would entail some postprocessing, see the last part of this message.

(A) Coloring applies across the whole value range, not just the unmasked part… e.g. even if you have masked everything above 6 A with some solid color (let’s say dark blue) and specified showing short-long distances with a white-black gradient, white is 0 angstroms and black is longest residue-residue distance in the protein, typically >50 A.  So the unmasked part of 0-6 A is only white to very light gray and you can’t really see any color differences…not that helpful if you are only interested in smaller differences between these shorter-range contacts.  (I realize you understand this problem, I’m just laying it out in case anybody else is following along.)  This problem is mentioned in the technical notes at the bottom of the help page for this tool:
<http://www.rbvi.ucsf.edu/chimera/docs/ContributedSoftware/rrdistmaps/rrdistmaps.html>

(B) Cannot show two RR Dist Maps dialogs at the same time.  So to see two maps side by side, you’d have to either run two Chimera sessions concurrently or within a single session, do one calculation, save image, do another, save image again or compare current dialog with previously saved image.

One way to compare the below-6-angstrom contacts only is to show side-by-side images of the distance map calculated for each protein separately with the distances >6 A masked out.  However, this will only show differences of where there is a contact <6A in one protein and not the other, but because of the coloring issue, subtleties in different distances within that range will not be perceptible.

Another way is to compare the individual distance maps (or a single map of average distance) with a difference map.  However, it would be difficult to tell exactly where the unmasked parts of the distance maps would be on the difference map and only look at those latter areas of the difference map.

(C) the Export button gives all the values of RR distance, and (if a comparison) standard deviation in RR distance and difference in RR distance.  There isn’t an option to give only the residue IDs for contact pairs within some distance; it only gives value matrices for all pairs.

(D) RR DIst Maps visualization is also focused on those 3 quantities, not attribute values.  Chimera attributes are assigned to individual atoms, residues, or models, not a pair of residues. The RR distance map shows residue-pair values with color.

If you want to use Chimera, for your purposes I think the best approach would be to use "Find Clashes/Contacts” or the “findclash” command to measure (intraprotein) CA-CA distances, and then filter to list only the pairs within 6 A.  Then you could do that for the other protein as well, then compare the two lists.  For example, if you had one of your proteins already open as #0, command:

findclash #0 at ca test self overlap -2.3 save ~/Desktop/protein1.txt 

<http://www.cgl.ucsf.edu/chimera/docs/UsersGuide/midas/findclash.html>

… would save a text file listing all CA-CA pairs with VDW surfaces within 2.3 angstroms.  The distance from atom center to atom center will be larger, but the output file will also give that distance and you can delete the pairs for which it is >6 angstroms.  I chose -2.3 because that should give you all the pairs within 6 angstroms and only a few with larger distances, using the VDW radii of CA atoms in structures without hydrogen atoms.  You may need to use a larger distance to get them all if your protein has hydrogens, but in any case, you will be able to tell by looking at the center-center distances included in the output.  The list is sorted by decreasing overlap so that longer distances willl be near the end.  This example command will automatically exclude residues that are adjacent in sequence (separated by no more than 4 bonds), but you could add the findclash command option “bondsep 2” to include those as well. 

You can use “log true” instead of (or in addition to) the “save” option to show the output in the Reply Log.  For example, if I have PDB 2gbp open as model #0, command:

findclash #0 at ca test self overlap -2.3 log true

… sends the following stuff to the Reply Log, primarily a list of the two atoms’ identifiers, atom-atom VDW overlap (negative value = separation between VDW surfaces of the two atoms), and atom-atom center-to-center distance:
-------------------
Allowed overlap: -2.3
H-bond overlap reduction: 0.4
Ignore contacts between atoms separated by 4 bonds or less
Detect intra-residue contacts: False
Detect intra-molecule contacts: True

620 contacts
atom1  atom2  overlap  distance
ARG 158.A CA  GLY 116.A CA  -0.203  3.963
LYS 189.A CA  GLY 217.A CA  -0.321  4.081
ARG 292.A CA  THR 110.A CA  -0.372  4.132
GLY 297.A CA  VAL 254.A CA  -0.453  4.213
THR 180.A CA  LYS 147.A CA  -0.460  4.220
VAL 235.A CA  ASN 211.A CA  -0.480  4.240
GLY 120.A CA  VAL 162.A CA  -0.493  4.253
LEU 196.A CA  ALA 201.A CA  -0.572  4.332
GLY 116.A CA  VAL 162.A CA  -0.663  4.423
MET 182.A CA  GLY 148.A CA  -0.668  4.428
THR 185.A CA  GLY 217.A CA  -0.697  4.457
ASN 256.A CA  ASP 236.A CA  -0.732  4.492
MET 182.A CA  GLU 149.A CA  -0.748  4.508
GLY 234.A CA  ASP 212.A CA  -0.794  4.554
[…. many lines …]
PHE 266.A CA  LYS 270.A CA  -2.269  6.029
ALA 155.A CA  THR 159.A CA  -2.274  6.034
ASN 302.A CA  GLU 305.A CA  -2.277  6.037
ILE 9.A CA    SER 41.A CA   -2.283  6.043
VAL 206.A CA  ILE 204.A CA  -2.285  6.045
LYS 92.A CA   LEU 67.A CA   -2.286  6.046
ALA 264.A CA  LEU 268.A CA  -2.286  6.046
LYS 270.A CA  ASP 274.A CA  -2.288  6.048
PRO 86.A CA   ALA 105.A CA  -2.289  6.049
ASN 130.A CA  HIS 126.A CA  -2.292  6.052
PHE 89.A CA   TYR 106.A CA  -2.293  6.053
ALA 251.A CA  VAL 245.A CA  -2.298  6.058
620 contacts
---------------
So, you would need to do some minor postprocessing to remove pairs with distances >6 angstroms and probably the harder part would be figuring out how to compare the two lists. Sorry I couldn’t give a more convenient solution. 
Elaine
-----
Elaine C. Meng, Ph.D.                       
UCSF Computer Graphics Lab (Chimera team) and Babbitt Lab
Department of Pharmaceutical Chemistry
University of California, San Francisco

> On Mar 22, 2017, at 12:44 PM, Nicole Persky <persky at broadinstitute.org> wrote:
> 
> Hi All, 
> I've just found the RRdistance Map function on Chimera and
> would like to compare between structures.
> I would like to:
> 1- make contact map of protein 1 to protein 1 with a cutoff of about 6 angstroms
> 2- make a contact map of protein 2 to protein 2 with a cutoff of about 6 angstroms
> 3- compare these two contact maps using the colors available for the whole contact map.
> 
> I can only seem to compare the ENTIRE contact map of protein 1 to the ENTIRE map of protein 2, whereas I'd like to mask any contacts over 6 angstroms.
> 
> Alternatively, if anyone knows of a way to export residue lists of contacts from the contact map site I could just compare them that way.
> 
> ALSO:
> Anyone know of a way to map defined attributes onto a contact map?
> 
> Thanks so much!
> Sincerely,
> Nicky




More information about the Chimera-users mailing list