*** For details on how to be removed from this list visit the *** *** CCP4 home page http://www.ccp4.ac.uk ***
Sometimes it helps to take a less automated approach. Make your Pattersons using a wedge of data between 10 and 5 Å resolution and see if you get the same sites. Since your native data extends to 1.8 Å, you should be doing phase extension with solvent flattening. Look at the maps before and after flattening to see if there is an improvement. Only when you are completely satisfied you have the right heavy atom sites and that the phasing is meaningful should you start to model density. At that point you can use partial model phase combination to further improve the map, but as soon as you start building into the density, you introduce the potential for model bias. Once you have your phases you can forget about the heavy atoms and use NCS along with solvent flattening to further improve the quality of the map (and the phase set). All that matters is that your native data have the local NCS. I seem to get the best solvent flattening results in CNS. Not clear why. Vineet Gaur wrote: > hi all > first of all i apologise for asking a non CCP4 question. > i m trying to solve a structure by MIR using Solve/Resolve. i have native > data up to 1.8A but the data quality for derivatives (Pt, Hg, Au and Sm) > is > not very food in terms of resolution and occupancies. due to this i have > calculated patterson at 3.3A with a mean fig of merits: 0.43. i m also > using > anomalous data for patterson calculation. Mean FOM after density > modification using resolve at 3.7A is 0.67. i cant further improve the > FOM using NCS although there r two molecules in asu, becz no NCS relation > exist between the heavy atom sites detected. since i m working at very low > resolution, automated model building is placing either ala or gly. Is > there > anything that i can try to further improve theFOM n secondly if i cant > further improve the FOM, how should i proceed in solving structure with a > partially build model for main chain. > thanx in advance > vineet gaur >
