Hi,<br><br>a free set in phenix.refine is choosen such that<br><br>- you have at least n% of free flags, but not more then 2000+epsilon (typically)<br>- it follows any possible twin law (not user specified, it follows symmetry of the lattice)<br>
<br>the ccp4 suite doesn't follow the same rules, so you will<br>- have more free flags<br>- have twin related reflections in your mixed between the test and work set. This reduces the rfree greatly and should be avoided.<br>
<br>P<br><br><br><br><br><div class="gmail_quote">2009/10/7 lauren <span dir="ltr"><<a href="mailto:lauren@gate.sinica.edu.tw">lauren@gate.sinica.edu.tw</a>></span><br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
<div link="blue" vlink="purple" lang="ZH-TW">
<div>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">Hi everyone,</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">I have a P31 twin crystal, resolution 2.0, and with</span></font><span lang="EN-US"> </span><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">twin fraction 0.451. I use phenix.refine to refine the
structure and with freeR set to 8%.</span></font></p>
<pre><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></pre><pre><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">% phenix.refine data.sca model.pdb twin_law="h,-h-k,-l" xray_data.r_free_flags.generate=True xray_data.r_free_flags.fraction=0.08</span></font></pre>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">Result:</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">Final: r_work = 0.1793 r_free = 0.2258 bonds = 0.011
angles = 1.557</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">REMARK�� 3�� FREE R
VALUE��������������������
: 0.2258</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">REMARK�� 3�� FREE R VALUE TEST SET
SIZE�� (%) : 3.67</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">REMARK�� 3�� FREE R VALUE TEST SET
COUNT����� : 2013</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">After the refinement, it seems like my freeR set size reduced
by half (3.67%). Is it due to the twin effect, or something is wrong with my
refinement?</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">And if I use ccp4 program �scalpeak2mtz � to
convert the .sca to .mtz and with freeR fraction set to 8%. Then refine it with
phenix.</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<pre><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">% phenix.refine data.mtz model.pdb twin_law="h,-h-k,-l" </span></font></pre>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">Final: r_work = 0.1669 r_free = 0.1970 bonds = 0.010
angles = 1.406</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">REMARK�� 3�� FREE R
VALUE��������������������
: 0.2022</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">REMARK�� 3�� FREE R VALUE TEST SET
SIZE�� (%) : 7.80</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">REMARK�� 3�� FREE R VALUE TEST SET
COUNT����� : 4326</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">It turns out, my test set is still 8%, and R/Rfree looks
even better than the first one. So I am wondering which one is the correct one</span></font><span lang="EN-US"> and why</span><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">?</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<pre><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">Thanks!</span></font></pre>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">Lauren</span></font></p>
<p><font face="Arial" size="1"><span style="font-size: 9pt; font-family: Arial;" lang="EN-US">�</span></font></p>
</div>
</div>
<br>_______________________________________________<br>
phenixbb mailing list<br>
<a href="mailto:phenixbb@phenix-online.org">phenixbb@phenix-online.org</a><br>
<a href="http://www.phenix-online.org/mailman/listinfo/phenixbb" target="_blank">http://www.phenix-online.org/mailman/listinfo/phenixbb</a><br>
<br></blockquote></div><br><br clear="all"><br>-- <br>-----------------------------------------------------------------<br>P.H. Zwart<br>Beamline Scientist<br>Berkeley Center for Structural Biology<br>Lawrence Berkeley National Laboratories<br>
1 Cyclotron Road, Berkeley, CA-94703, USA<br>Cell: 510 289 9246<br>BCSB: � � <a href="http://bcsb.als.lbl.gov">http://bcsb.als.lbl.gov</a><br>PHENIX: <a href="http://www.phenix-online.org">http://www.phenix-online.org</a><br>
CCTBX: �<a href="http://cctbx.sf.net">http://cctbx.sf.net</a><br>-----------------------------------------------------------------<br>