t

INTERNATIONAL SUMMER SCHOOL ON MACROMOLECULAR CRYSTALLOGRAPHIC COMPUTING

FINAL REPORT

Recipient Organization: Department of Energy

A

Prepared Under Grant No. DE-FG03-96ER62226

GENEHL AIOIHICS

DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency tbereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or use- fulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any spe- cific commercial product, process, or scMcc by trade name, trademark, manufac- turer, or otherwise does not necessarily constitute or imply its endorsement, m m - mendation. or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

DISCLAIMER

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

GA- A22 94 9 RECE'vQ) Page 1 of 3

A& 2 4 199 Report on the 1996 International U n i e 4 yys!allography (IUCr)

Computing Schoo

Summary

The International Union of Crystallography (IUCr) Macromolecular Crystallography Computing School was held at Western Washington University in Bellingham, Washington between August 17 and 22, 1996. The School was the seventh in a series of IUCr Crystallographic Symposia. There were 106 attendees from 16 countries; of these 38 were either speakers or tutors. Of the remaining 68 attendees, 29 were graduate students, 12 were postdoctoral fellows, 9 were faculty, and 18 were from industry or elsewhere. The format of the School was formal lectures in the morning, tutorials in the afternoon, and software demonstrations and more lectures in the evening.

Draft proceedings were distributed to attendees of the School and final papers appear in the published proceedings which is being distributed by Oxford University Press. For the first time at a Computing School full Proceedings are available electronically via the World Wide Web (http://www.sdsc.edu/XtaVIUCr/CC/School96/IUCr.html) or via anonymous ftp from ftp.sdsc.edu in the directory /pub/sdsc/societiesmTCr/School96.

The full program which left both the organizers and attendees exhausted, reflects the current state of excitement in the field of macromolecular structure determination using the technique of X-ray crystallography. Our understanding of biological processes has gained immeasurably from work already done, and as shown in Figure 1 of the paper presented by Stewart, we are climbing rapidly along an exponential growth curve that promises many more major discoveries before the next IUCr Congress in 1999. It is estimated that at the 1999 Congress the approximately 5,000 structures solved at the time of the 1996 School will have grown to over 20,000. The new and improved technologies and techniques described in these Proceedings are contributing to that growth and at the same time, as pointed out in the paper given by Sussman, creating challenges for the Protein Data Bank (PDB).

As the School progressed, we were struck by the similarities to events which took place in small molecule crystallography beginning some 20 to 25 years ago. Growth then was fueled by the advent of new algorithms, affordable computer hardware, and good software. So it is today for macromolecular crystallography, but with the added bonus of the Internet which is changing how we conduct our research. Hack presented this view as part of his on-going contribution to how crystallographers use the Internet.

As the number of available structures increase, thoughts turn to relative accuracy and common and different features between structures. The relative ease with which the majority of structures are solved today, effects how we work and how we think about each individual structure and structures en masse. Thornton and Wodak in there papers focused on comparative analysis of proteins with regard to accuracy, and Thornton, Bryant, and Bourne considered the problem of classification and comparison of proteins. With so many structures, databases are becoming vital weapons in the crystallographers armory. Gilliland , Sussman, and John Westbrook described their work with the crystallization database, 3DB, and the Nucleic Acid Database, respectively. Good databases require a complete and computer usable representation and hence Fitzgerald and John Westbrook provided an overview of the macromolecular Crystallographic Information File (mmCIF) and associated tools, respectively.

After presentations discussing structures en masse we returned to the more traditional mode of presentation which parallels the determination of a single macromolecular structure: data collection - phasing- model building and visualization - refinement.

GA- A2 2 9 4 9 Page 2 of 3

Howard provided attendees with an overview of the steps in the collection of macromolecular data. This was followed by a presentation from Minor discussing recent advances and the even more important issue of accuracy. Related to accuracy (and sanity) is the development of graphical user interfaces (GUIs) which reduce mistakes when collecting data. Sweet described the GUI they have developed for the X12-C line at Brookhaven National Laboratory. Finally Mary Westbrook described the software system being developed for data acquisition at the Structural Biology Center (SBC), located at Argonne National Laboratory's Advanced Photon Source, which was just coming on-line at the time of her presentation.

The discussion on phasing techniques for macromolecules began with the Plenary lecture delivered by Hauptman who explained how least squares can lead to a formulation of the minimal principle and effectively replaces the phase problem by one of constrained global optimization. The Shake-and-Bake software addresses this reformulation and was the subject of a presentation by Weeks who highlighted the successful application of Shake-and-Bake in ab initio phasing of proteins of 600 non-hydrogen atoms or less where data to greater than 1.2 8, are available. Hauptman went on to describe how the use of single wavelength anomalous scattering (SAS) diffraction data leads to yet another formulation in global optimization and hence solution to the phase problem. Giacovazzo took a more optimistic view of the power of direct methods as applied to phasing protein structures, and went as far as to suggest that direct methods is competitive with single isomorphous replacement techniques. If low resolution phases are available, Main described how they could be used in a rotation function yielding better results that the standard rotation function that operates in vector space. Also on the topic of phase extension Collins described work using an objective function and Bricogne described maximum likelihood heavy atom parameter refinement for the MIR and MAD methods.

On the subject of model building and visualization, a paper delivered by Edwards on behalf of Oldfield described the procedures available in QUANTA 96 for the semi-automated fitting of electron density maps, a time consuming step in the structure solution process. Finzel described the use of existing structural information for map fitting in the LORE program and Nicholls described the representation and the beginnings of the comparative analysis of protein surfaces. Features embodied in the program GRASP.

On the subject of refinement Sheldrick described the use of SHELXL in the refinement of proteins using high resolution data. He touched upon issues of constraints versus restraints, restrained anisotropic refinement, the role of Rfiee, disorder, the solvent model, radius of convergence and esd's. Ten Eyck described the role of full matrix least squares refinement for macromolecules while Bricogne described the use of maximum likelihood as opposed to least squares in refinement using a combination of the programs BUSTER and TNT. Chapman returned us to the use of real space refinement, fust pioneered by Diamond in 1971, but now using stereochemical restraints in the refinement of large virus structures and fitting know structures to low resolution electron microscope reconstructions. Finally Tronrud introduced the notion of joint refinement using Thermolysin inhibitors. Restraints are constructed based on regions of the multiple structures which are invariant versus those regions which are variant - an increasingly common occurrence.

The latest developments in software used by many of us were described by their authors. McRee described Xtalview, Steigemann described the PROTEIN system, concentrating on molecular replacement and density modification in real space, Dodson described the CCP-4 package, Turk described MAIN, and Sheldrick described the application of SHELX to macromolecules.

Finally, a session was devoted to either new approaches or ancillary techniques. In the former category Johnson described work in crystallographic topology, and in the latter Moss described the application of object oriented programming to crystallography, Wild described an AVS interface to the CCP-4 programs, and Ramanadham described the paralielization of crystallographic codes.

These Computing Schools remain the most important function of the IUCr Computing Commission in educating young crystallographers in the foundations of computational crystallography and in also bringing us not-so-young crystallographers up to speed on the latest developments. This school was not meant to train crystallographers in the use of normally available crystallographic software, but to introduce them to the latest developments in the field and what new directions are being pursued. It is our hope that seeds of

GA-A22949 Page 3 of 3

new ideas were planted in the minds of the attendees and that the students of this school will be the tutors and lecturers of the subsequent schools presenting exciting new developments.

Acknowledgments

The School would not have been possible without generous sponsored by the San Diego Supercomputer Center (SDSC), the Collaborative Computational Project for Protein Crystallography (CCP4), Digital Equipment Corp., Merck Research Laboratories, Molecular Simulations Inc. (MSI), the National Science Foundation (NSF), Reciprocal Space Ventures Inc., Silicon Graphics Inc., and the US Department of Energy.

On a personal note thanks to Angela Loh for arranging the loan of a variety of Digital hardware and software, and to Jim Aquino and Vaughn Winslow, also of Digital, for getting it in place. Thanks to Monte Montebello of Silicon Graphics for arranging the loan of SGI hardware and software.

Thanks to the Western Washington University staff - Diane Bakkom, Manager of Conference Services, Lois Longwood, also from Conference Services, Susan Komsky, Director of Academic Technology, Gerry Price and Terry Dayton from Computer Services, and especially Jerry Huddle from media maintenance, who sat through every session, and learnt more about computational crystallography that perhaps he wanted.

Thanks to John Westbrook who helped reduce our level of panic in managing all the loaned computer systems.

A special thanks to the SDSC folks. he-meeting Gerry Nordstrom and Jim D’Aoust for taking care of the accounting and especially to Holly Boyce who took care of everything, including the organizers! Finally, in terms of the published and electronic Proceedings thanks to Gail Bamber for her artwork and especially to David Hart for copyediting the whole volume.

Internet Access Statistics Access to the 1996 Computing School Web pages (http://www.sdsc.edu/XtaVIUCr/CC/School96/IUCr.html) are as follows. The statistics are for 1997. Since the pages were officially announced in late March of 1997 access is for a nine month period.

There were a total of 27,090 Web hits in 1997.

The majority of papers were available only in Postscript via ftp. Statistics are more limited from the ftp logs. There were 1721 downloads in the period Sept 1 1 - Dec 22, 1997 - ie 17/day well after the initial announcement period.

Most papers had at least 10 hits. It may be that some people systematically downloaded all files.