A data set has been collected on a crystal of protein X at cryo temperature. The data were collected by Arie Schouten, beginning of January 1996, at the University of Utrecht (The Netherlands, Crystal- and Structural Chemistry) with a DIP2000 detector. The spindle axis comes from the left seen from the direction of the beam, and lies with the beam in the horizontal plane.
The unit cell of the crystal is 73.9 73.9 200.9 90 90 90, space group 
(or 
).
A copper rotating anode was used (
), the crystal to detector
distance was 165 mm. No theta offset was used: x beam 99.95 y beam 99.93.
After indexing, refinement and integration of the first frame, STRATEGY was run with the COMPleteness 98 96 94 92 90 option. The global minimal oscillation ranges for the different total completenesses are given in the following table.
Table: Minimal oscillation ranges for different completenesses for datacollection on protein X
Figure 
shows the minimal needed oscillation range as a function of
the starting spindle angle for a 92% complete data set in the resolution range of
20 - 2.9
. There is a minimum between 23 and 30 degrees. Starting at these spindle
angles, the total sweep needed for a data set of at maximum 92% completeness is a bit
less than 40 degrees.
In total 80 frames of 0.5 degree oscillation were collected, spanning an
oscillation range of 40 degrees and starting at a spindle angle of 24 degrees.
Complete dataprocessing of these frames (indexing, refinent, integration, scaling
and rejection of the bad merging reflections) using
DENZO and SCALEPACK gave the following redundancy table:
Shell Summary of observation redundancies by shells:
Lower Upper No. of reflections with given No. of observations
limit limit 0 1 2 3 4 5-6 7-8 9-12 13-19 >19 total
20.00 6.19 278 232 269 188 173 238 59 0 0 0 1159
6.19 4.94 179 213 250 197 215 239 46 0 0 0 1160
4.94 4.32 143 203 259 234 213 232 32 0 0 0 1173
4.32 3.93 109 189 249 277 233 222 19 0 0 0 1189
3.93 3.65 94 185 278 284 238 196 8 0 0 0 1189
3.65 3.44 101 186 277 287 274 161 3 0 0 0 1188
3.44 3.26 67 204 274 346 230 147 0 0 0 0 1201
3.26 3.12 89 214 309 303 222 129 0 0 0 0 1177
3.12 3.00 92 222 309 318 226 97 0 0 0 0 1172
3.00 2.90 104 231 322 293 220 73 0 0 0 0 1139
All hkl 1256 2079 2796 2727 2244 1734 167 0 0 0 11747
Table of redundancy and completeness for a data collection of 80 frames,
starting at the spindle angle 24.00 and with an oscillation range of 0.50 degree(s).
Shell Summary of predicted redundancies by shells:
Lower Upper No. of reflections with given No. of predictions
limit limit 0 1 2 3 4 5-6 7-8 9-12 13-19 >19 completeness
20.00 6.19 247 176 288 187 178 266 95 0 0 0 82.81
6.19 4.94 155 160 255 203 204 278 85 0 0 0 88.43
4.94 4.32 123 149 258 232 211 289 55 0 0 0 90.66
4.32 3.93 92 136 244 267 243 283 32 0 0 0 92.91
3.93 3.65 74 134 246 289 251 269 19 0 0 0 94.23
3.65 3.44 69 123 248 308 277 256 6 0 0 0 94.64
3.44 3.26 47 113 237 330 295 248 2 0 0 0 96.31
3.26 3.12 57 113 222 335 314 219 1 0 0 0 95.48
3.12 3.00 42 111 212 334 356 205 0 0 0 0 96.67
3.00 2.90 42 91 216 329 392 180 0 0 0 0 96.64
All hkl 948 1306 2426 2814 2721 2493 295 0 0 0 92.71
The pattern of this table corresponds well with the prediction
of STRATEGY, based on the first integrated
frame with the NFRAme 80 command.
STRATEGY predicts more reflections than have actually been measured.
This is because of rejection of reflections by DENZO due to bad background,
bad profile fitting, rejection of partial measured reflections due to the mosaicity
(
, rejection of spatially overlapping reflections and/or rejection of bad merging
intensities. Especially the number of
overlaps can greatly influence the discrepancy between the predicted and the observed redundacy table
(however, there were hardly any overlapping reflections in these frames).
Furthermore, most of the parameters predicting the diffraction patterns will be
continuously refined during the integration with DENZO, which makes them more accurately defined
than the parameters obtained from the first integrated frame that were used by STRATEGY.
In order to guess what the influence of the bad background or bad profile fitting and
the (post)refinement is, 80 empty frames have been integrated in the same oscillation range.
The same mosaicity of 0.5 was used. After changing all the intensities to the arbitrary
value of 100, the following redundancy table was produced using SCALEPACK:
Shell Summary of observation redundancies by shells:
Lower Upper No. of reflections with given No. of observations
limit limit 0 1 2 3 4 5-6 7-8 9-12 13-19 >19 total
20.00 6.19 251 181 290 189 176 260 90 0 0 0 1186
6.19 4.94 159 168 258 209 197 267 81 0 0 0 1180
4.94 4.32 126 160 255 228 214 287 46 0 0 0 1190
4.32 3.93 98 139 252 268 237 276 28 0 0 0 1200
3.93 3.65 80 142 250 295 244 258 14 0 0 0 1203
3.65 3.44 76 144 255 297 279 234 4 0 0 0 1213
3.44 3.26 53 140 284 309 272 208 2 0 0 0 1215
3.26 3.12 67 180 268 302 267 182 0 0 0 0 1199
3.12 3.00 72 173 276 313 268 162 0 0 0 0 1192
3.00 2.90 73 179 268 327 259 137 0 0 0 0 1170
All hkl 1055 1606 2656 2737 2413 2271 265 0 0 0 11948
This table is pretty similar to the prediction of STRATEGY, which shows
that neglecting the mosaic spread by STRATEGY is allowed for our purpose.