Dear Almudena, I have some questions to help you, how many molecules are presents in the ASU ? If more than one, maybe you can try to find NCS in the substructure. To perform that it could be necessary to increase (more than 20) the radius of NCS research parameters. You can try to find NCS with sitcom. Or by hand. If you can use NCS to average density during the solvent flatening steps, it could help.
If you haven't NCS in your substructure. After phenix and solvent flattening step, is the map interpretable ? Is phaser keeping all the sites fund by HySS ? or it find less sites ? Can you see "nice" density, continue, protein shape ? or is it more or less a lot of blob random ? You can try to select only the best sites found by HySS, maybe some are not well determined. HySS locate 127 but how many is it suposed to find ? One last thing, for the autobuild programm it's not the same to work with dataset at 1.8 A of res compare with one at 3 A. And you must not expect the same level of quality. And if you look the model build, are you able to continue manually, correct some clearly wrong chain? In other word be as sure as possible that the initial phases that you give to autobuild programm are not to far from the good one. Hope to help you. I am not very familiar with HySS that's why i don't go deeper for this parts. But you can try shelx which is parts of CCP4 or via HKL2map. And compare the sites obtain with these from HySS. Nicolas ________________________________________ De : CCP4 bulletin board [CCP4BB@JISCMAIL.AC.UK] de la part de Almudena Ponce Salvatierra [maps.fa...@gmail.com] Envoyé : vendredi 26 juin 2015 12:32 À : CCP4BB@JISCMAIL.AC.UK Objet : [ccp4bb] anomalous phasing with HySS - phaser - Autobuild Hi all, I have SeMet data for which HySs locates 127 Selenium atoms with a CC of 0.31, which I think is decent. Then I run Phaser to generate the first maps, and it gives a score of 25 and LLG of 190 or so... The next step would be running Autobuild, however, the first models I realize it is building have R work and R free between 50 and 55%. So I guess it is just some random stuff there. If the result from HySS and Phaser looks more or less good, why does AutoBuild show these numbers?? Any idea of what I could do to improve so? Any ideas are welcome. Best wishes, Almudena -- Almudena Ponce-Salvatierra Macromolecular crystallography and Nucleic acid chemistry Max Planck Institute for Biophysical Chemistry Am Fassberg 11 37077 Göttingen Germany
