Notes from meeting held at Daresbury Laboratory, 25-26th July 2001
A meeting took place at the Daresbury Laboratory involving the main people presently responsible for initial implementation of the expert system.
The aim of the meeting was to discuss and resolve technical details where possible, to review progress to date and ensure further progress of the project. The meeting started on the afternoon of the 25th and ended at 15:00 on the 26th.
Business for this meeting
Review of progress
XML definitions
There was a fairly lengthy discussion of the XML associated with COLLECT commands to the data collection software and INDEX commands to the MOSFLM server. The outcome is the following
<collect_start> <data-set id = "1"> <resolution="2.0"> <exposure_time="1.5"> <phi_start="0.0"> <phi_end="30.0"> <oscillation_range="1.0"> <no_passes="1"> [<slit>] <wavelength="0.978"> <directory="/data/dna"> <filename_template="steve_###.img"> <first_image_number="1"> </data_set> </collect_start> <collect_output> <status> <data_set_id> <error_message> </collect_output> <collect_end> </collect_end> <index_start> <project id1="ls1402" id2="run 4432" [id3,id4...]/> <fileinfo directory="/data/id14eh4/external/ls1402/se1/e1" filename_template="bfft_###.img" /> <image no="001" [phi_start="0.0"]/> [<image no="031" [phi_start="6.0"]/>] [<image no="061" [phi_start="12.0"]/>] [<detector type="adsc" gain="1.0" x_pixelsize="0.0816" y_pixelsize="0.0816"/>] [<beamline divergence="0.001" polarisation="0.98" /> ] [<data_collection wavelength="0.8374" distance="223.2" phi_range="0.2" x_beam="95.45" y_beam="93.43"/>] [<cell a="74.4 b="38.2" c="38.6" alpha="90.0" beta="90.0" gamma="90.0"/>] [<symmetry sym="p212121"/>] </index_start> <index_output program="mosflm" > <project id1="ls1402" id2="run 4432" [id3,id4...]/> <status flag = "ok" /> <solution no_reflections_used = "2342" no_reflections_rejected = "123" rms_spot_deviation = "0.234" x_beam_updated = "95.55" y_beam_updated = "93.23" laue_group = "P" [penalty = "10.0"]/> [<symmetry sym="p212121" />] <cell a="74.4 b="38.2" c="38.6" alpha="90.0" beta="90.0" gamma="90.0"/> <orientation_matrix type="amat" program="mosflm" e11="0.00328538" .../> /> </index_output>Communication between the expert system and data collection programs
The flow of control between the various process involved in the overall data collection was discussed. It was felt the data collection GUI should remain the single point of user interface wherever possible. So the GUI would perhaps have a button 'SCREEN CRYSTAL', this would then initiate a communication with the expert system.
In actual fact the expert system would probably the ask for some images to be collected. Once completed the expert system would be informed, or perhaps it should be told images are ready as they are saved to disk. The question was raised here as to whether we can know this for certain with the supplied manufacturer software for ALL detectors. Once images were complete the expert system would initiate a communication with the data processing software, resulting in one or more solutions.
The expert system could then either pick a solution, according to criteria we need to define, or pass all solutions back to the data collection GUI. In the latter case the data collection GUI could leave the choice up to the user, then passed back to the expert system. Finally (for now anyway) the expert system would get the data processing software to determine the best strategy for data collection, and request the data collection software perform it.
The process can be presented graphically as follows. The diagram ignores the MOSFLM server as a separate entity for simplicity.
Commands necessary for running MOSFLM to achieve indexing
Two sets of commands to achieve indexing have been suggested.
One as generated by Elves.
#
GENFILE auto.gen
#
# Detector parameters (range, quantum gain, etc.)
SCANNER omega 90 ROTATION horiz ANTI FAST horiz ORIGIN UR RECTANG TYPE ADSC
OVERLOAD CUTOFF 60000
GAIN 0.3
LIMITS RMAX 133 YMAX 188 XMAX 188 xscan 188 yscan 188
# Beam parameters (i.e. (delta-lambda)/lambda, H,V divervence in degrees)
WAVELENGTH 0.932998
DISPERSION 0.0025
DIVERGENCE 0 0
POLARISATION SYNCHROTRON 0.9
# Alignment parameters
DISTANCE 183.11
BEAM 94.75 93.80
TWOTHETA 0.0
#
NEWMATRIX auto.mat
# Crystal parameters
#CELL KEEP 100 100 100 90 90 90
#SYMMETRY P1
RESOLUTION 1.5
MOSAIC 0.15
# frame filename breakdown
DIRECTORY /mntdirect/_scisoft/pxsoft/data/bfr_21/images
IDENTIFIER bfr_21
EXTENSION img
#
RESOLUTION 3.0
AUTOINDEX DPS IMAGE 1 [45 71 90]
#
GO
Another as generated by DPS.
TITLE lys_1, frames 1 to 30
GENF lys_1.gen_1_30
EXTENSION img
DIRE /net/pxfs3/data1/ccd_test/Lysozyme/
HKLOUT lys_1_1_30.mtz
IDENT lys_1
SYMMETRY p43212
MATRIX lys_1.mat
NEWMAT lys_1.mat
MOSAIC 0.500
IMAGE lys_1_001 phi 31.000 to 32.000000
!SCANNER ADSC
SCANNER ROTATION HORIZ ANTI FAST horizontal ORIGIN UR RECT TYPE ADSC
SIZE 2304 2304
PIXEL 0.0816 0.0816
FINDSPOTS YOFFSET 4.0 THRESHOLD 20.0
TIMEOUT 2
LIMITS RMIN 5 RMAX 134.3 XMAX 95.0 YMAX 94.9 XSCAN 94.0 YSCAN 94.0
BEAM 95.020 93.060
BACKSTOP RADIUS 5 CENTRE 95.020 93.060
ADCOFFSET 20
NULLPIX 0
GAIN 0.300
DISTORTION TILT 0.000000 TWIST 0.000000
DISTANCE 120.500
!beam parameter
SYNCHROTRON POLAR 0.89
WAVELENGTH 0.9790
DIVE 0.030 0.010
DISPER 0.0025
!integration parameters
SEPARATION 0.600 0.600 CLOSE
RASTER 15 15 6 1 1
RESOLUTION 1.70
REFINEMENT CYCLES 2 RESID 15 IMIN 20 5 NSIG 3
REFINEMENT INCLUDE PARTIALS 0.5
REFINEMENT LIMIT 30.0
REFINEMENT FIX YSCALE
OVERLOAD NOVER 0 CUTOFF 55000
PROFILE XLINES 0 47 94 141 189 YLINES 0 47 94 141 189
GO
Set up standard test data sets
Two 'standard' data sets were exchanged. One of lysozyme and one as follows
BFR_21: filename_template = "bfr_21_###.img"
start_phi = "180"
no_images = "90"
oscillation_range = "0.3"
distance = "183.11"
x_beam = "94.75"
y_beam = "93.80"
two_theta = "0.0"
wavelength = "0.9329661"
Summary of achievements during the meeting
Aims prior to next full meeting
The aims prior to the next full meeting were defined more clearly as follows.
Business identified for the next full DNA meeting
Actions