VMD-L Mailing List

From: John Stone (johns_at_ks.uiuc.edu)
Date: Fri May 22 2009 - 10:57:39 CDT

Hi,
  It calculates the average position, then computes the variance for
each atom position in the selection, then takes the square root:

// Calculate RMS fluctuation of selected atoms over selected frames
extern int measure_rmsf(AtomSel *sel, MoleculeList *mlist,
                        int start, int end, int step, float *rmsf) {
  if (!sel) return MEASURE_ERR_NOSEL;
  if (sel->num_atoms == 0) return MEASURE_ERR_NOATOMS;

  Molecule *mymol = mlist->mol_from_id(sel->molid());
  int maxframes = mymol->numframes();

  // accept value of -1 meaning "all" frames
  if (end == -1)
    end = maxframes-1;

  if (maxframes == 0 || start < 0 || start > end ||
      end >= maxframes || step <= 0)
    return MEASURE_ERR_BADFRAMERANGE;

  int i;
  for (i=0; i<sel->selected; i++)
    rmsf[i] = 0.0f;

  int rc;
  float *avpos = new float[3*sel->selected];
  rc = measure_avpos(sel, mlist, start, end, step, avpos);

  if (rc != MEASURE_NOERR) {
    delete [] avpos;
    return rc;
  }

  // calculate per-atom variance here
  int frame, avcount, j;
  for (avcount=0,frame=start; frame<=end; avcount++,frame+=step) {
    const float *framepos = (mymol->get_frame(frame))->pos;
    for (i=0, j=0; i<sel->num_atoms; i++) {
      if (sel->on[i]) {
        rmsf[j] += distance2(&avpos[3*j], &framepos[3*i]);
        j++;
      }
    }
  }

  float avinv = 1.0f / (float) avcount;
  for (j=0; j<sel->selected; j++) {
    rmsf[j] = sqrtf(rmsf[j] * avinv);
  }

  delete [] avpos;
  return MEASURE_NOERR;
}

On Thu, May 21, 2009 at 12:14:17AM -0400, Maria Sanchez wrote:
> Hi All,
>
> I want to learn how VMD calculates the RMSF. From a previous post, I know
> that for the frames and atom selected, it calculates an average position of
> the atom for the specified frames. I presume that then it calculates the
> distance between the average position and the position each time. But,
> my calculation doesn't match the result that I obtain using the rmsf
> function. Can somebody explain me how vmd does this calculation?
>
> Thanks,
> Maria Antonieta 

-- 
NIH Resource for Macromolecular Modeling and Bioinformatics
Beckman Institute for Advanced Science and Technology
University of Illinois, 405 N. Mathews Ave, Urbana, IL 61801
Email: johns_at_ks.uiuc.edu                 Phone: 217-244-3349
  WWW: http://www.ks.uiuc.edu/~johns/      Fax: 217-244-6078