<html>
  <head>
    <meta content="text/html; charset=windows-1252"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    Hi Peter,<br>
    <br>
    you suggest to calculate "the (free) likelihood".. May I ask:
    specifically likelihood of what you are suggesting to calculate and
    what is "free" in this context? I guess I'm just lost in jargon,
    sorry!<br>
    <br>
    Thanks,<br>
    Pavel<br>
    <br>
    <br>
    <div class="moz-cite-prefix">On 2/11/15 8:08 PM, Peter Zwart wrote:<br>
    </div>
    <blockquote
cite="mid:CALE9nFGyyxYA5=AOvF7+fscgANwqDTPzCOFLwE=T0fmCGUqahQ@mail.gmail.com"
      type="cite">
      <div dir="ltr">
        <div>Dear All,</div>
        <div><br>
        </div>
        How meaningful are the second derivative based estimates
        obtained via full matrix inversion when the gradient is not 0
        (i.e. when not in the minimum)? I can understand that when you
        are working with high-resolution data and your R-value is close
        to 0, things could work, but what happens when around a more
        challenging 2A?�
        <div><br>
          <div>If you are interested in the uncertainty of the
            occupancy, I recommend not doing any refinement, but just
            generate a list of occupancies and B-values for the atom of
            interest and compute the (free) likelihood for each model.
            Subsequent normalisation of the neg-exponent of these
            values, should provide you with an answer that could be just
            as believable as any other method around.� A little bit of
            python scripting should do the trick quite easily.</div>
          <div><br>
          </div>
          <div>Both the full matrix inversion and the suggestion above
            probe the steepness of the data-agreement hole the structure
            is sitting in. Pavels suggestion explores the spread of
            local minima around the starting configuration. I am not
            sure what method is more appropriate, perhaps it is
            instructive to know what problem you are trying to solve.</div>
          <div><br>
          </div>
          <div>HTH</div>
          <div>P</div>
          <div>�</div>
          <div><br>
          </div>
          <div><br>
          </div>
          <div><br>
          </div>
        </div>
      </div>
      <div class="gmail_extra"><br>
        <div class="gmail_quote">On 11 February 2015 at 16:13, Masaki
          UNNO <span dir="ltr">&lt;<a moz-do-not-send="true"
              href="mailto:unno19@mx.ibaraki.ac.jp" target="_blank">unno19@mx.ibaraki.ac.jp</a>&gt;</span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0
            .8ex;border-left:1px #ccc solid;padding-left:1ex">Dear all<br>
            <br>
            Thank you very much for your suggestions.<br>
            I will try making a number of models in which the atom has
            different<br>
            occupancies (e.g. 0.1-1.0). Then, I will refine them by
            restraining the<br>
            B-factors.<br>
            Actually, our structure contains some reaction intermediates
            not only the<br>
            substrate. So I would like to estimate the ratio.<br>
            <br>
            Best regards<br>
            <span class="HOEnZb"><font color="#888888"><br>
                Masaki<br>
              </font></span><span class="im HOEnZb">-----Original
              Message-----<br>
              From: <a moz-do-not-send="true"
                href="mailto:phenixbb-bounces@phenix-online.org">phenixbb-bounces@phenix-online.org</a><br>
              [mailto:<a moz-do-not-send="true"
                href="mailto:phenixbb-bounces@phenix-online.org">phenixbb-bounces@phenix-online.org</a>]
              On Behalf Of Pavel Afonine<br>
            </span><span class="im HOEnZb">Sent: Thursday, February 12,
              2015 6:36 AM<br>
              To: Dale Tronrud; <a moz-do-not-send="true"
                href="mailto:phenixbb@phenix-online.org">phenixbb@phenix-online.org</a><br>
              Subject: Re: [phenixbb] How should we estimate the
              "uncertainty" of the<br>
              occupancy of an atom?<br>
              <br>
            </span>
            <div class="HOEnZb">
              <div class="h5">Hi Dale,<br>
                <br>
                &gt; P.S. I'll look up the paper you reference but my
                university does not<br>
                &gt; subscribe to acta Cryst and getting those papers
                takes time.<br>
                <br>
                it is open access:<br>
                <br>
                <a moz-do-not-send="true"
                  href="http://phenix-online.org/papers/wd5073_reprint.pdf"
                  target="_blank">http://phenix-online.org/papers/wd5073_reprint.pdf</a><br>
                <br>
                All the best,<br>
                Pavel<br>
                <br>
                <br>
                <br>
                _______________________________________________<br>
                phenixbb mailing list<br>
                <a moz-do-not-send="true"
                  href="mailto:phenixbb@phenix-online.org">phenixbb@phenix-online.org</a><br>
                <a moz-do-not-send="true"
                  href="http://phenix-online.org/mailman/listinfo/phenixbb"
                  target="_blank">http://phenix-online.org/mailman/listinfo/phenixbb</a><br>
                <br>
                <br>
                _______________________________________________<br>
                phenixbb mailing list<br>
                <a moz-do-not-send="true"
                  href="mailto:phenixbb@phenix-online.org">phenixbb@phenix-online.org</a><br>
                <a moz-do-not-send="true"
                  href="http://phenix-online.org/mailman/listinfo/phenixbb"
                  target="_blank">http://phenix-online.org/mailman/listinfo/phenixbb</a><br>
              </div>
            </div>
          </blockquote>
        </div>
        <br>
        <br clear="all">
        <div><br>
        </div>
        -- <br>
        <div class="gmail_signature">
          <div dir="ltr">
            <div>-----------------------------------------------------------------<br>
              P.H. Zwart<br>
              Staff Scientist<br>
              Berkeley Center for Structural Biology, Science lead<br>
              Lawrence Berkeley National Laboratories<br>
              1 Cyclotron Road, Berkeley, CA-94703, USA<br>
              Cell: 510 289 9246<br>
              SASTBX:� <a moz-do-not-send="true"
                href="http://sastbx.als.lbl.gov" target="_blank">http://sastbx.als.lbl.gov</a></div>
            <div>BCSB:� � � <a moz-do-not-send="true"
                href="http://bcsb.als.lbl.gov" target="_blank">http://bcsb.als.lbl.gov</a><br>
            </div>
            <div>PHENIX:�� <a moz-do-not-send="true"
                href="http://www.phenix-online.org" target="_blank">http://www.phenix-online.org</a></div>
            <div>-----------------------------------------------------------------</div>
          </div>
        </div>
      </div>
      <br>
      <fieldset class="mimeAttachmentHeader"></fieldset>
      <br>
      <pre wrap="">_______________________________________________
phenixbb mailing list
<a class="moz-txt-link-abbreviated" href="mailto:phenixbb@phenix-online.org">phenixbb@phenix-online.org</a>
<a class="moz-txt-link-freetext" href="http://phenix-online.org/mailman/listinfo/phenixbb">http://phenix-online.org/mailman/listinfo/phenixbb</a>
</pre>
    </blockquote>
    <br>
  </body>
</html>