Hi all,<br>thanks to several tremendously helpful replies I got the molecular replacement, which I bugged the ccp4bb with during the last weeks, under control. Thanks for all the input, I&#39;ll send out a summary soon.<br>
<br>One question remains, though:<br><br>With an improved search model, the molecular replacement in Phaser got much stronger. This time, the correct rotation solution was on top of the list, and translation Z-scores were &gt;10. ARP/wARP could not build into these maps, even after cross crystal averaging. In the end, no utter surprise given the low homology of the search model with my protein of interest. 
<br>Molrep also succeeded with this model . After the standard sequence rotation/translation/rigid body, Molrep reported R-factors of 31% and 41% for data up to 4AA or 3AA, respectively. However, when I used the Molrep output for a rigid body refinement in Refmac, R-factors were at 55% for 4AA data.
<br><br>This huge gap of R-factors from Molrep and Refmac is really puzzling to me.<br>Any ideas?<br><br>Cheers<br>Jan<br><br>