Integrated system for macromolecular crystallography data collection and analysis at X9B beamline at NSLS

The goal of the newly developed data collection and processing system is to integrate all computational activities that have to be performed during the diffraction data collection experiment. The HKL-2000_X9B system that evolved from the HKL-2000 package has been implemented on X9B beamline at the National Synchrotron Light Source (NSLS) in the Brookhaven National Laboratory. Efficient data analysis changes the protein crystallography experiment as the data collection is a part of a process that includes indexing, integration, scaling and phasing. In this approach, the final result of such an experiment is a high quality electron density map that provides assurance that the experiment was successful. The expected result of efficient data collection/processing is not only high-throughput crystallography but also reduction of the effort needed to produce proteins and grow crystals of sufficient quality for structure solution.

The whole hardware used for macromolecular crystallography experiments, apart from the beamline optics, is completely autonomous and can be moved to another station within hours. There are only two connections to that autonomous system: the power and Ethernet cables.

The HKL-2000_X9B system is based on a client-server concept. The proprietary for X9B beamline server hklserver_x9b interacts with all beam line hardware: kinematic table, goniostat, detector, beamstop, monochromator, cryogenic system, and hutch safety devices. The hklserver_x9b is a multi process that runs on a dedicated computer and provides a central point that controls the exchange of messages between the beam line hardware and all software installed on the beam line computers.

The server contains the following objects:

  • Kinematic table: horizontal and vertical translations and rotations

  • goniostat motion system: w,k,j-movements and detector distance

  • crystal alignment motion system: x,y,z-movements

  • detector lift

  • beamstop motion system: x,y,z-movements

  • slits motion system

  • X-ray shutter

  • energy (wavelength) of the X-ray beam

  • fluorescence spectrum recordin

  • beam intensity monitoring (ionization chambers)

  • hutch door status

  • crystal freezing system

The Graphical User Interface (GUI) controlls all stages of the experiment:

  • table and beam optimization

  • crystal evaluation and alignment

  • set-up of the experimental parameters

  • execution of data collection

  • progress monitoring

  • data processing

  • data archiving

Monitoring the data collection process includes the same features as the software for crystal evaluation and is available for any data image at any time. In addition, the monitoring software has the capability to compare crystal properties at different times during the crystal lifetime. The status of the system is represented graphically by the real-time updated OpenGL made in scale model of goniostat together with detector and slits system. The collision map was constructed and takes into account the collimator, detector distance, w and k goniostat positions.

The crystal centering is performed in semi-automatic mode that requires the experimenter to indicate the current crystal position which moves the crystal to the marked beam center position. There are two pre-programmed goniostat positions: zero goniostat and crystal recovery. Crystal recovery position is suitable for fast mounting and dismounting of the crystals contained in the vials filed with liquid nitrogen.

The first version of HKL2000_X9B has been installed together with new hardware in January 2002. The initial installation and commissioning time took 5 days. During the first week of operation, 3 MAD structures have been solved, one during the data collection session. Between January and September 2002, in three sessions of 3-4 days each, several HKL2000_X9B options described above were implemented and tested. During the first days of September 2002, the test data on the crystal of tetragonal lysozyme were collected using a wavelength of 1.54 λ. This experiment was analogous to the previous one (Dauter et al., 1999) except that the data were collected in a single pass in 1.5 hours instead of previous four passes to 1.55 λ resolution requiring the whole day. The current data extended only to 2.03 λ and above 60 of the strongest reflections were lost due to the overloaded profile pixels. Nevertheless, SHELXD identified 17 anomalous scatterers (ten sulfurs and seven chloride ions) and SHELXE ( produced a very clear electron density map. The whole procedure, from data scaling and merging (with HKL2000_X9B), through phasing (with SHELXD and SHELXE) and map calculation and display, took about half hour after the last image has been recorded. The collection of the 1.54 λ wavelength data and the subsequent solution of the structure of lysozyme can be therefore proposed as a convenient test for the commissioning of the new synchrotron beamlines.

The increase of the Internet bandwidth and in particular the arrival of the Internet-2 will provide an opportunity to interact remotely with the experimental setup and to perform a synchrotron experiment from a home laboratory, although data collection through the Internet will require the automation of the sample mounting. The high level of abstraction implemented in HKL-2000_X9B will allow it to serve as a prototype for the systems controlling other synchrotron beamlines as well as the home X-ray equipment.

People Involved