March 2004 NSLS-II Workshop, Crystallography Breakout Session

The present Memo summarizes the topics discussed by the Macromolecular Crystallography user community during the Macromolecular Crystallography Breakout Session at the NSLS-II workshop held March 15, 2004 at Brookhaven National Laboratory. Following a brief introduction, by Vivian Stojanoff, in which features of NSLS-II were described, the session was divided into two parts; in the first, chaired by Larry Shapiro of Columbia University, three scientific talks summarized specific concerns of the user community in regards to science, instrumentation, and user issues:

Wayne Hendrickson, "Challenges in Biological Crystallography"
Paula Fitzgerald, "Practical Experiences in Running a Shared Ownership Beamline"
Elizabeth Duke, "Macromolecular Crystallography and Diamond - Exciting Prospects with a Third Generation Source"

In the second part of the session, a panel moderated by Leemor Joshua-Tor (Cold Spring Harbor Laboratory), and consisting of Chris Lima (Sloan-Kettering Institute), Wayne Hendrickson (Columbia University), Lonny Berman (NSLS), Larry Shapiro (Columbia University), Paula Fitzgerald (Merck and Co.) and Elizabeth Duke (Diamond Light Source), addressed questions from the audience.

Main recommendations from the speakers and the following panel discussion addressed:

Facility and beamlines
Infrastructure for biology
Quality-of-life
Locality and commercial users
Other issues (funding)

Facility/Beamlines

A general consensus was reached that macromolecular crystallography beamlines should be preferentially located on insertion devices. One major concern was the number of proposed beamlines (5) on insertion devices, as was described during the morning plenary session; at the NSLS-II macromolecular crystallography workshop hosted by Cold Spring Harbor Laboratory (Banbury) and held last summer, "a consensus was reached in that at least 7 insertion devices should be dedicated to this research area" (taken from the workshop summary). The newly proposed number (5) might be acceptable if "chicaning" within the straight sections is permitted, which could allow for 5 straight sections to accommodate 10 insertion devices (2 for each straight section) to serve as radiation sources for 10 independent beamlines. It was pointed out by Wayne Hendrickson that at current facilities and for proposed facilities, 1/3 of the beamlines are typically reserved for macromolecular crystallography.

A main concern with the proposed storage ring was the time frame in which it will become operational. It was questioned if an energy recovery linac based machine would not serve the community in better ways. Since the technology is not completely understood and in place for these machines it was recommended that the current proposal should be easily upgradeable once the technology has been proven.

The environment around the beamlines was brought up by Chris Lima, in general agreement with the audience. The beamline controls and space for setting up should be enclosed for privacy and comfort, as is the existing paradigm at the ESRF. Beamline operation in the open experimental hall, as is currently conducted at the NSLS and at most beamlines at the APS, was agreed to contribute to the stress of conducting an experiment at a synchrotron radiation facility.

Another concern was about potential use of the floor space for a "sector" that views a straight section and adjacent bending magnet, in the event that it is decided to build out more than one insertion device beamline in the sector, with each insertion device beamline operating independently of the other(s) and viewing separate insertion devices whose radiation emissions do not overlap (as a result of "chicaning" as pointed out above). Would there be sufficient floor space to do so, and avoid encroachment by one or the other (or both) insertion device beamlines on the floor space set aside for an adjacent bending magnet beamline? Would sufficient space remain for control areas, especially if it is desired that such control areas should be enclosed? Clear answers to this depend not just on the total available floor space that would be brought to bear, but also on the detailed beamline designs (e.g. forward-directed as in the case of GM/CA-CAT at the APS vs. sideways-deflected as in the case of ID14 at the ESRF). These issues apply not just to the macromolecular crystallography community, but also to all user communities of NSLS-II. It is suggested that, at a very early stage of planning, a working group of interested users, beamline staff, and persons with expertise in conventional construction and in ESH issues should be formed to study these issues and make recommendations, well before conventional construction plans for the experimental floor are far advanced.

End stations should be a mixture of fixed set-up and flexible set-ups. Special end stations will probably be science and scientist-in-charge driven. No consensus was achieved on needs of special instrumentation for spectroscopy measurements, lasers, pressure environments, and magnetic fields. Development of new detector technologies to be compatible with the brightness and flux as well as sample damage was not much discussed but considered essential to the progress of structural biology as performed at these high brightness sources. Wayne Hendrickson pointed out the recent published finding that, at the level of the hottest beamlines that are currently available at the APS, damage to macromolecular crystals inflicted by radiation is still dependent on the total absorbed dose, not the instantaneous dose rate. Assuming that this trend continues to extrapolate to the level of photon flux that will be available at NSLS-II beamlines, the primary challenge will be able to collect the data faster than presently feasible. Doing so will require advances in detectors, goniometers, and shutters as well (unless "shutter-less" detection such as with pixel-array detectors can be developed), not to mention data transfer technology and automated sample exchange to minimize down-times between successive data collections.

Wayne Hendrickson also presented some nice data clearly exhibiting the enhanced phasing power that is afforded by collecting MAD data at higher incident beam energy resolution. As such, he suggests that an effective strategy for "attenuating" the incident beam intensity (if it happens to be too strong to permit successful data collection, owing to technical shortcomings such as those described above), would be to narrow the incident beam energy resolution through the use of slits or higher index monochromator crystal Bragg reflections, resulting in both attenuated intensities as well as enhanced MAD phasing power potential at the same time.

Special attention was called by Wayne Hendrickson, as well as by Rod MacKinnon in his plenary session lecture, to the importance of having very small incident beams (~50 µm) and small beam stops, to enhance the signal-to-background ratio and permit the ability to measure very low resolution data (around the beam stop shadow) as well as discriminate between adjacent Bragg reflections from very large unit cell crystals. Rod, during his lecture, presented quantitative comparisons to highlight the importance of micro focus incident beams (citing data collected at ESRF), but also remarked that a very bright x-ray source such as NSLS-II is needed to make usable beams of this dimension possible. The expected NSLS-II beamline performance data which Vivian Stojanoff presented in her introduction, and which are described in the submitted NSLS-II proposal, show that the incident beam sizes will be ~50 µm at the sample position, and therefore it is necessary to construct appropriate apertures and collimators to deliver clean beams of this dimension, and incorporate motorized sample alignment stages and high-resolution sample viewing capabilities to take fullest advantage of this small beam size.

While the majority of macromolecular crystallography beamlines for NSLS-II might be preferred to be of the "generic" variety (featuring optimum spectral performance in the range of 10-15 keV), Elizabeth Duke mentioned that a later phase beamline for Diamond, that will be optimized for longer wavelengths (to ~2.5 Ang, i.e. ~5 keV) is under discussion, for the benefit of single wavelength data collection to maximize the signal from anomalous diffraction from sulfur atoms. Elizabeth Duke also cautioned about the need to increase the radiation shielding that is necessary for an insertion device beamline, due to bremsstrahlung radiation hazards that increase as the straight section is lengthened. The optics enclosures that will be used on such beamlines at Diamond, for example, require transverse bremsstrahlung radiation scatter shielding of lead thickness 30 mm to surround the entire enclosures.

Robotics and automation requirements are a given. This model is being followed by Diamond which is part of the DNA collaborative effort. The main argument to follow with totally automated beamline control and experimental set-up is the diverse nature of the user community and the increasing numbers of users with little or no experience in the field. Mail -in services and remote access are also considered as a must for the new facility.

Elizabeth Duke also noted that some users are returning to room temperature data collection on extremely fragile samples grown in capillaries. For these kinds of samples, it is especially helpful to incorporate sample viewing along two different directions on the diffractometer, with one of this on-axis (i.e. in-line with the incident beam path); the latter is already being pursued at several beamlines at the ESRF and the ALS for example.

Infrastructure for Biology

In general the "Biology Village" model being proposed for the Diamond facility, in which beamlines for biology research are clustered at one section of the ring, seems to be very attractive. It would be cost effective, allow better interaction between users and staff besides allowing for shared centralized support. One example is the relatively large number of cold rooms currently available at the NSLS and other facilities, which could be reduced significantly. It should also be considered to place other life sciences beamlines in close proximity with the proposed structural biology beamlines.

Laboratory space seemed to be appropriate in the draft NSLS-II floor plan, but it is not clear from the current proposal if this space would also include storage space. Storage is particularly important for beamlines with flexible experimental set-ups. If the current area proposed also includes storage, then it might not be large enough.

The beamlines must be well staffed to permit their active use around-the-clock. Paula Fitzgerald pointed out, however, that the adopted model for PRTs and CATs in the past, to employ high-level scientific and engineering staff during the design, construction and commissioning stages isn't the best, as there occurs little use for such expertise after operation commences. She points out that it is better that such expertise is provided by the facility, with perhaps contractual arrangements made with the facility for such expertise during the construction and commissioning phases. A similar model was adopted by the ESRF, and it has been proven essential that the scientist in charge of the beamline operation was involved throughout the design, construction, and commissioning stages.

Quality-of-Life

This was not really discussed during the formal session, but in the coffee and lunch/dinner breaks. A major complaint of the life sciences community focuses on the current housing conditions at BNL. The modern, capacious Guest House model pursued by the APS was praised, both in relation to the quality of the housing service as well as its proximity to the facility. This model is currently being followed by the SSRL. It was pointed out that such a facility even right now would significantly benefit the present NSLS. Cheap and efficient transportation is also an issue. Users are either stranded on-site over the weekend, having to provide their own meals on a Saturday night, or depend on a car. Cafeteria operation hours have been a large concern of our user community for years, not only because the reduction in the hours of operation but as well the quality of the food being served. Again models such as the ones found at the APS and at the ESRF were praised and recommended for immediate consideration.

Locality and Commercial Users

Wayne Hendrickson pointed out that the strength of the current facility is mostly in the technological development that is being done, laying the groundwork for most of the third generation facilities available now. This should again be a major concern of the proposed NSLS-II. In Wayne's words: "technical challenges for the NSLS-II can be divided into: fundamental -- radiation damage and limited crystalline order; and practical -- small crystals, large unit cells, low-resolution structures, mega-site sub-structures and high-throughput." Paula Fitzgerald also argued that, for pharmaceutical users, proximity would not be a strong selling point of the proposed facility, but rapid access and efficiency would be "better sellers", especially when upheld with automation and mail-in services. However, it was argued by Rod MacKinnon, in his plenary session lecture that proximity is of vital importance for users who are pursuing challenging projects and absolutely require hands-on access. For example, frequent hands-on access to a bright x-ray source to provide constant feedback between the synchrotron and the biochemistry lab is essential when crystal growth is being optimized for difficult systems such as membrane proteins and very large complexes. In such cases, automation and mail-in services are of less value.

No consensus was reached in regards to the participation of commercial users towards the beamlines. The model of shared and individual facilities was discussed. Small commercial users seem to prefer a collaborative model such as available to them through the existing CRADA or SBIR system.

Other Issues

The question, How to provide funding to build and operate beamlines and supporting facilities, was a major concern of the Banbury meeting that led up to the current NSLS-II proposal and the March 15 workshop. Even though there is a consensus in the community that the beamlines should be built and operated by the NSLS-II facility on a collaborative basis, it is not clear how this model can fit with any funding scheme. It is clear that this model cannot be followed if raising the funding is left to individual groups. An interesting model was proposed by Paula Fitzgerald: the model followed by the members of the IMCA facility at APS. In this model, 12 companies formed a consortium, and each bought a share and started to pay dues to build up a construction budget well before the actual construction of the IMCA-CAT sector at the APS had started. A similar organization could fit with the proposed user access at the NSLS and a collaborative beamline construction and operation model sought by the macromolecular community. As pointed out by Paula: it is essential to start planning now in order to be ready when the NSLS-II ring is turned on.

It is the recommendation of the organizers of this workshop to organize another workshop to address specifics of beamline and end station configuration with the participation of different members of the macromolecular crystallography synchrotron facilities. The purpose of the workshop would be to determine the needs and design aspects of the macromolecular crystallography beamlines. The resulting document could be used to approach funding agencies and industrial partners. It is also proposed to address the funding agencies with a project in the style of the DOE BER Genomes to Life program.

Last Modified: January 31, 2008
Please forward all questions about this site to: Gary Schroeder