|
1992 Media Coverage
- December, 1992:An article in the President's Column of Impact discusses the
directed methods NIH project. It states that the relationship between
the Medical Foundation of Buffalo and
Dr. Russ Miller
has already produced algorithms for structure
determination substantially more effective and powerful than existing
programs. It goes on to discuss
the fruitful relationship as well as the impact of a parallel computing
solution to the problem targeted at the Connection Machine and Intel iPSC
parallel supercomputers.
- November, 1992:An article in The Science Teacher covers
the spectacular success of the Shake-and-Bake research team, led by
Dr. Russ Miller
and Nobel Laureate Herbert A. Hauptman, in terms of the overnight success
of two previously unresolved and difficult molecular structures.
- October 15, 1992: An article appears in the
Reporter
entitled "Crystal Problem Solver: Computer method yields speedy solution"
that states a computer scientist
(Dr. Russ Miller)
with no crystallographic experience has adapted
and applied an automated method to solve the unknown crystal structures of
two molecules literally overnight.
The long article states that these structures had been worked on for over
10 years before Miller got his hands on the data and used the Shake-and-Bake
algorithm and him parallel computing solution to solve the structures overnight.
The article states that the breakthrough was announced August 10 in Pittsburgh
at the American Crystallography Meeting. Nobelist Herbert A. Hauptman states
that "the fact that someone who is not a crystallographic expert
could just run this computer program like a black box and get a correct
solution for these two unknown and admittedly difficult structures is
really impressive." The article goes on to quote Miller as explaining
the new computer program and the new algorithmic strategy that the team
of Miller, Hauptman, and other members of MFB developed. The solutions
were determined on a Thinking Machines CM-5, a parallel computer housed
at Thinking Machines Corp.
- August 26, 1992: An article appears in Inside R & D that
is currently unavailable.
- Fall, 1992: An article in FNSM Focus mentions that
Dr. Russ Miller
has accepted an invitation to join the executive committee of the newly
formed IEEE Technical Committee on Parallel Processing and that his
responsibilities will include heading the effort in undergraduate
education in parallel processing. The article also mentions that
Miller has adapted and applied an automated method to solve
the unknown crystal structures of two molecules. The solution to each of these
(as yet unnamed) structures has eluded
crystallographers for more than 10 years. Finally, the article states
that Miller will serve on teh program committee of Frontiers `92: The
Fourth Symposium on Frontiers of Massively Parallel Computation.
- Summer, 1992: An enormous article in UB Research that
takes 2 of the 6 pages of the publication covers the "Molecular
Shake-and-Bake: New technique solves structure of large molecule
overnight, duplicating work that previously took years."
This article describes the shake-and-bake method of molecular
structure determination that was developed by Dr. Russ Miller and numerous colleagues at the
Medical Foundation of Buffalo.
It describes the solution of a previously solved structure, called Gramicidin
A, that had originally taken 10 years to solve. Shake-and-Bake solved the
structure overnight. Nobel Laureate Herbert A. Hauptman stated that
"its the firs time that a structure of this complexity has been solved
by this technique." The article goes on to describe the myriad applications
of the shake-and-bake method, potentially to cancer, heart disease,
thyroid disorder, arthritis, diabetes, hypertension, stroke and AIDS.
The article quotes Miller as saying that "our algorithm takes as input
tens of thousands of random molecules, generated by the computer and from
each produces a potential solution. We need to generate enough potential
solutions so that with high probability, we are able to guarantee that one
of them represents the correct molecule." Further,
Miller explained that using Hauptman's minimal function and basic
chemical knowledge, the positions of the atoms in the randomly generated
molecule are moved about hundreds or thousands of times until they
stabilize and produce a solution. Miller points out that
"Hauptman's function not only guides us in the generation of potential
solutions, but it is also used as a predictor in selecting the most
likely of the potential structures. Finally, the article discusses
parallel computing implementations of the algorithm.
- September, 1992: An article in Impact states that
Dr. Russ Miller
has joined the Molecular Biophysics Department of the Medical Foundation of Buffalo, Inc. (MFB) The article goes on to discuss the fact that Dr. Miller will spend
his sabbatical year working with Drs. Hauptman and Weeks on the development
of algorithms that exploit a parallel computer environment to solve the
phase problem of x-ray crystallography, which is the basis for determining
the molecular structures of biologically important materials. Further,
the article states that Dr. Miller's collaboration with the MFB team has
already succeeded in that the group solved a three hundred atom structure
on two different parallel architectures, which is significantly larger than
a molecule that can typically be solved by these so-call direct methods.
Dr. Miller is quoted as saying that "we need massive amounts of computer power,
and the only way to get power today is with some sort of massively
parallel computer. Miller noted that the group did the calculations on
time borrowed on an NIH machine and on a machine at Thinking Machines Corporation,
for which Miller consults. The article concludes with a short biography of
Miller.
- August 27, 1992: An article in the Reporter states that Dr. Russ Miller
is author of "Multisearch Techniques for Implementing Data Structures
on a Mesh-Connected Computer," which will appear in the Journal of Parallel
and Distributed Computing. The article is co-authored by M.J. Atallah and
J.-J. Tsay, Purdue, F. Dehne and A. Rau-Chaplin, Carleton University.
- August 23, 1992: An article in the
Buffalo News states
that a complex structural problem that had stumped crystallographers
for the past 10 years was wrapped up overnight by
computer scientist Dr. Russ Miller, when he ran it for 90 minutes on a large parallel supercomputer.
- August 22, 1992: An article appears in Science News
entitled "Shaking and baking to atomic positions" states that an expert
crystallographer can take literally years to determine the positions of
atoms in a molecule. However, using a
new technique developed in Buffalo, it appears that the time can be
reduced from days to hours. The article gives an excellent and detailed
description of x-ray crystallography and molecular structure determination.
It then describes the minimal function developed several years ago by
Dr. Herbert A. Hauptman, Nobel Laureate and President of the Hauptman-Woodward
Medical Research Institute. It then goes on to state that "to make the
scheme workable, computer scientist Russ Miller of SUNY-Buffalo and his co-workers developed
algorithms specifically designed to run efficiently on computers consisting
of large numbers of parallel processors, which operate simultaneously while
working on different pieces of the problem." The article then describes the
algorithm in some detail and quotes Miller as saying that "you essentially
toss the atoms in a shoebox with enough chemical knowledge to make sure the
arrangement makes chemical sense" before migrating back and forth between
a set of atoms and a set of the phases that they represent, making minor
adjustments or perturbations in each space. Once this process is complete
on a number of such initial sets of starting atoms, Hauptman's function
can be used on the resulting sets of atoms to determine which ones are
likely solutions. In describing the solution to one important and complex
structure Miller is quoted as saying that "I put the data into the program
before I went to bed at night and the next morning I looked through the
results" only to find that one of the 64 random arrangements migrated to
a set of atoms that, according to Hauptman's function, represented a solution.
Miller turned over the information to David Langs who refined the structure
into a complete solution. After this success, Langs provided Miller with
a data set of another difficult and unsolved structure, which Miller was
also able to solve quite rapidly.
- August 13, 1992: An article in the
Buffalo News describes some of the
work that Dr. Russ Miller
is involved with at the Medical Foundation of Buffalo. The article
discusses new breakthroughs (Shake-and-Bake) and potential applications
of the technology. It also mentions the recent $3 million grant from NIH.
The article quotes Dr. Miller as saying that "if you
know the structure of a drug, you can deliberately improve it,
you can understand why it does what it does, and
you can figure out if using it causes long-term problems."
- May/June 1992: An article appears in Computers in Physics,
vol. 6, no. 3, entitled "Minimal Principle Facilitates Structure Determination
of large Molecule." The article describes the pioneering work of
Nobel Laureates Herbert A. Hauptman and Jerome Karle and the fundamentals of
x-ray crystallography. The 3-page article then goes on to describe the new
Shake-and-Bake method of molecular structure determination and quotes
Dr. Russ Miller
as stating that "we need massive amounts of computer power and the only
way to get power today is with some sort of massively parallel computer."
The article goes on to state that Miller, who was instrumental in writing the
computer code, says that the program has been run on an Intel iPSC/860
at NIH and on a Thinking Machines Corporation CM-2 with 16,000 processors
and on a Thinking Machines Corporation CM-5. The article goes on to
discuss the team's design, development, implementation and application of
the procedure to large structures, including a previously solved 317 atom
molecule (Gramicidin A), which originally took 14 years to solve and was
solved overnight with Shake-and-Bake using the original data. The article
concludes with a positive outlook in that Shake-and-Bake could represent
the future of a black-box approach to solving molecular crystal structures,
a key step in rational drug design.
- May 13, 1992: An article appears in The Chronicle of Higher Education that is currently unavailable.
- May 7, 1992: An article appears in the Reporter stating that Dr. Russ Miller was awarded a Research Experiences for
Undergraduates Supplement from the National Science Foundation to support
the work of Steve Gallo, who will work with Dr. Miller on the phase
problem of x-ray crystallography.
- April 25, 1992: An article appears in the
Buffalo News stating that
a Buffalo-based research team led by a Nobel laureate will unveil a new
"shake-and-bake" research method that could speed the development of
drugs targeting a wide range of diseases. The technique, which uses a
combination of parallel supercomputers and human insight, cuts the time involved in "solving" the molecular structures of complex chemical compounds from years to just a few weeks.
For example, the article points out that David Langs of the Medical Foundation
of Buffalo once managed to analyze the molecular structure of Gramicidin-A,
a bacteria-derived antibiotic. It took him 10 years. Using the new technique,
Dr. Russ Miller was
able to use the software developed by the team to solve the structure overnight.
This new method, which took nearly three years to develop, will be unveiled in
Florida during a meeting of crystallographers.
"It's the first time that a structure of this complexity has been solved by this technique," said Hauptman, leader of the team and co-winner of the 1985 Nobel Prize in chemistry.
The article also mentions that last month,
the foundation's Research Institute won a $3 million, four-year federal grant that officials said may make it the key center worldwide for the study of ways to determine the structure of the body's basic construction materials and viruses that attack the body.
The computer program -- which runs on parallel supercomputers -- generated
hundreds of thousands of random molecules, which scientists then narrowed down
to a smaller sample that they thought was most likely to lead to a solution.
The programs then "move" the individual atoms in the
mathematically modeled molecules hundreds or thousands of times, until they stabilize and produce potential solutions.
- April 17, 1992: An article in Science
entitled "New Methods Make Mid-Sized Molecules Easier to Solve" states
that it is typically quite difficult to solve structures in the important
200-500 atom range. After a lengthy discussion of existing methods and
their failures in that size range, the article discusses the newly
developed Shake-and-Bake method of molecular structure determination
and their promising results when testing the procedure on Gramicidin A
on a Thinking Machines Corporation CM-5.
- April 17, 1992: An article in The Wall Street Journal entitled "Taking Trial-and-Error out of Designing
Drugs" states that for years, researchers have developed drugs largely
by manipulating molecules of simple chemicals and using trial-and-error
testing to pick the most effective configuration. Now, with "rational
drug design" the techniques have advanced. Further, the article goes on
to describe the importance of knowing the 3-D structure of a molecule and
then discusses the Shake-and-Bake algorithm. The article quotes
Dr. Russ Miller
as saying that "this will drastically reduce the time it takes to develop
a drug, from years to weeks." Dr. Miller also discusses the application and
ramifications of employing parallel processing techniques in terms of the
running time of the calculations.
- April, 1992: An article in Impact entitled
"Medical Foundation Gets $3 Million Grant" discusses the details of
an NIH grant that focuses on the phase problem of x-ray crystallography.
Scientists working on this project include Drs. Hauptman, Langs, Dorset,
Griffin, Weeks, Smith, Guo and Miller. The article also discusses the press conference on
February 24, when the grant was announced. In addition to staff from
MFB and UB was Congressman Henry Nowak. The title of the project
is "New Methods of Biomolecular Structure Determination."
- April, 1992: In this issue of "In the News: University at Buffalo
in the National News" is the April 17, 1992 article from the Wall Street entitled "Taking Trial-and-Error out of Designing
Drugs" (see above).
Journal
- March 26, 1992: An article in the Reporter discusses the $3 million of funding awarded by NIH
to a group of scientists from the Medical Foundation of Buffalo that includes
Dr. Russ Miller.
The article describes the seminal work of Nobel Laureate Herbert A. Hauptman,
President of the Medical Foundation, who heads the project. It also describes
the advantages of our proposed research in terms of attacking problems such
as disease in a more intelligent fashion. Dr. Miller is quoted as
saying that "the current solution involves generating large sets of
random molecules, and then applying an extensive and computationally
intensive 'recycling' strategy to each one with the expectation
that at least one will converge to the correct solution." Miller goes
on to say that "basically, one can think of the process as tossing a random
set of atoms into a shoebox and continually moving the atoms about until they
stabilize, while keeping in mind chemical information and Hauptman's minimal
function." The article reviews some of the successes the team has had to
date and describes some of the parallel machines that Miller has developed
the code for and the results on such machines, including the Connection Machine
CM-5.
- March 4, 1992: An article in the Buffalo News entitled "Medical Foundation gets $3 million grant: study
examines proteins, viruses," discusses the 4-year $3M grant awarded to the
Medical Foundation of Buffalo for work in protein structure determination.
This grant, which is the largest grant ever received by the Medical Foundation,
will focus on the Shake-and-Bake method of molecular structure determination
and involves numerous scientists at the Medical Foundation as well as
Dr. Russ Miller.
|