Director: Robert Glaeser, LBL; Wah Chiu,
Baylor College of Medicine
The Core Project (F) consists of two
sub-projects, both of which support the objectives of the other
projects with in the program as a whole. In Project F1
we propose
to acquire test-image data-sets for ice-embedded single particles,
which will be archived in a server that is available simultaneously to
the software developers in this program project and to the community
via the Internet. Two different specimens with known atomic
structures will be used in order to provide a reality check for
validating our computational methodology. The broader
availability of the data sets should further stimulate new algorithmic
development towards atomic structure determination by cryo-electron
microscopy of single particles. Our primary test molecules will
be the GroEL particles and tubulin (in the form of microtubules), both
of which are chosen because of their appropriate size and inherent
symmetry. Project F2 will provide key
scientific infrastructure
that is needed within the program. Project F2
will also provide
local scientific expertise in cryo-EM that is needed to support the
work of Projects C and D, for which the primary expertise of the PIs
lies in computer science and mathematics respectively. Finally,
Project F2 will coordinate a range of
centralized functions that are
related to efficient communication and collaboration between all
projects
The Core research Project has two budget-centers: Baylor College of
Medicine in Houston, and LBNL in Berkeley. The Houston center
(Project F1) will be responsible for
creating, maintaining and
distributing sets of single-particle reference-data for use in
single-particle software development. The LBNL center (Project
F2) will provide extensive scientific
coordination between all of the
separate projects, thereby providing centralized liaison for the
program project as a whole. The three specific aims of Project F are
(1) to create large sets of experimental, single-particle
reference-data, and maintain these data sets for distribution; (2) to
provide access to 80-node (Pentium III) cluster at LBNL for
beta-testing by software developers and by associated PIs (Project G)
within the program project, and to facilitate access to the IBM SP
supercomputers at NERSC when it's needed for software testing; and (3)
to provide backbone of scientific coordination within the program
project.
Project F1-Generation
of A High-Resolution Test Image Dataset of Single Particles
In Project F1, we plan to acquire complete datasets of images of
ice-embedded protein molecules, all of which have detectable signal
beyond 4Å. These images will be used as test data sets for
the validation of the computational methodologies (at resolutions
ranging from 4-10Å), which are to be developed in the program
project as a whole. One test molecule will be GroEL, for which
the X-ray structure is well known. A second test "molecule" will
be short segments of microtubules, for which a chain-trace structure
(of the tubulin dimer) has been derived by electron
crystallography.
The initial phase of this project will involve determining the best
specimen preparation protocols, the most efficient instrument and
recording medium for the intended project and the optimal imaging
conditions. Collaborative and remote interaction between the
Houston and Berkeley labs will be set up for investigators to work out
the optimal conditions for data collection by sharing simultaneous,
remote microscopic observations. The chosen instrument and
operational conditions will ultimately take advantage of automation for
routine, computer-assisted, high-throughput image data-collection.
The second phase of the project will involve recording data for these
two specimens. The digital images will be characterized by their
de-focus, signal-to-noise ratio (SNR) at various frequencies and
experimental B factors, determined in conjunction with X-ray scattering
or crystallographic data. The original micrographs and the
boxed-out particle images will be archived for easy access by other
investigators via the Internet. Along with the raw images, we
will include all the imaging conditions and sample preparation
descriptions. The data set will be simultaneously accessible to
all investigators in the program project and the outside
scientific community.
Project F2-Infrastructure
and Scientific Coordination
The purpose of Project F2 is to maintain scientific coordination and
communication within the program project, and to provide technical
liaison between projects within the program. Project F2 will
ensure that technologies developed by Projects C and D are incorporated
into the production packages developed by Projects A and B,
respectively. It will also test the portability of parallelized
software packages and new software modules and compare the speed and
performance of parallelized versions of software packages.
Furthermore, Project F2 will facilitate access to computational
infrastructure that is available at LBNL, and provide scientific and
technical liaison with (1) the PIs of the associated projects (Project
G) and (2) the cryo-EM community as a whole.
.
