Hi All, Thanks a lot for the advice from Mark.
1. For the comment to use a attenuated beam, the major concern here is resolution, since at full beam, the resolution is already somewhere between 3.5-4.2, I assume that significant attenuation will bring down the resolution dramatically, and also decrease the I/Sigma which might lead to more errors in |F|? 2. We used TaBr because as a large electron dense cluster, it is good for low resolution phasing. This derivative, in the best scenario, give a 4.8A data set (a resolution that is supposed to be enough to find the huge cluster) which only has 0.1~0.2A unit cell difference to a native data set collected on the same beam, theoretically, this dataset should give prominent peaks on hacker section (the crystal is green, and fluoresence scan give moderate signal), but what I find is only two small peaks (6-7 sigma, boarder line) (maybe occupancy too low?), and using Hyss to find the HA only result in a FOM of 0.20. This is what triggers me to wonder if a high scaling error model (0.05-0.1) is also detrimental for isomorphous phasing? 3. speaking of Se. Is it possible to use Se as a second derivative for MIR in my case (500 amino acid, 18 Met)? my concern is that Se might be too “light” for my protein, and 4.5A resolution is NOT much better than the 7A Mark has mentioned. 4. Has anybody ever used xenon? Can anyone share the starting pressure and time for optimization? Thank you again, Bei
