Q1. How do you create double mutants in yeast on a genome-wide scale? Isn't efficiency a problem? A1. We take advantage of the fact that
Saccharomyces cerevisiae
can assume both diploid and haploid states. Our
starting point is the
Yeast Knockout strain collection,
from which we have repurified, screened, and pooled all of the
available heterozygous diploid strains. To the single null mutation
in these strains, we add a second null mutation through a standard
gene replacement strategy involving homologous recombination.
The trick is to get a high enough transformation yield so that
each strain is adequately represented in the pool. Low
transformation yields (< 2 x 105 colonies) result in
noisy data. These protocols were optimized by Xuewen Pan and
further refined by Pam Meluh et al. The latest version is from
March 2007.
Q2. How are the heterozygous diploid double mutants turned into haploid double mutants? A2. Again, we take advantage of the well-known ability of
Saccharomyces cerevisiae to undergo meiosis through
sporulation. These protocols has been optimized for efficiency as
well. The latest version is from
December 2005.
Q3. How is cell viability in the double mutants assayed experimentally? A3. The sporulated plated cultures are plated separately onto
matched selective and non-selective growth media and grown for 48
hours. The cells are harvested by scraping, and genomic DNA is
extracted. The latest protocol is from
December 2005.
Q4. How do the microarrays assess cell viability? A4. They measure cell abundance via the relative abundance of
oligonucleotide "tags"
flanking each null mutation cassette. The tag sequences are
amplified by PCR through shared universal priming sites that were
also incorporated into the mutation cassette. (The primer-tag-primer
sequence is akin to a "barcode".) The PCR primers are labeled so
that the PCR products can be hybridized to microarrays in two-color
experiments.
Q1. Where can I get the Yeast Knockout strain collection? A1.The two vendors we are working with are
(Open Biosystems, Inc.)
and the American Type Culture Collection
(ATCC).
As of October 25, 2005, the heterozygous diploid mutants that we
have constructed and validated for our Project are now available
as the Yeast Magic Marker strain collection from
Open Biosystems.
These strains contain a "Magic Marker," an elaboration of the
MFA1-HIS3 selectable marker introduced by
Tong et al. (2001),
that enables selection of MATa haploid strains from a
sporulated culture.
Under a licensing agreement between Open Biosystems, Inc. and
the Johns Hopkins University, the University is entitled to a share
of royalties on sales of Yeast Knockout strains described in this
web site. The terms of this arrangement are being managed by the
Johns Hopkins University in accordance with its conflict of interest
policies.
Please note: SLAM project scientists receive no personal
compensation as a result of sales of these strains.Q2. Where can I get tag microarrays? A2. Our microarrays were custom-designed by us but are now
commercially available through Agilent Technologies, Inc.
As of July 28, 2005, a yeast barcode array consortium has been
established under the aegis of Princeton University. All researchers
are able to join the consortium in order to purchase the microarrays
from Agilent; however, only those affiliated with academic
institutions are eligible for the negotiated consortium price. To
join the consortium, please register at the
consortium website.
Please note: Neither the SLAM project nor its scientists
receive compensation of any kind as a result of sales of these
arrays.Q3. Where can I get a list of primers, solutions, etc. needed for these experiments? A3. We are putting together web-friendly versions of these
lists. Until then, please see the publications listed elsewhere.
Q1. How do you analyze data from your microarrays? A1.The data presented here have been analyzed by Brian
Peyser with ad hoc code written in the
R programming language,
which is widely used by academic statisticians and freely
downloadable on all major computing platforms. The
control/experimental log ratio data have been normalized according
to average log intensity. No background correction was performed.
UpTags and DnTags are combined through a calculation that is
described in a paper currently under review.
In addition, Daniel Yuan has written a software package called
hoptag that is designed to systematize the data
processing pipeline. This work is being done in collaboration with
Rafael Irizarry
in the Department of Biostatistics at the Johns
Hopkins Bloomberg School of Public Health. hoptag is a
lightweight adaptation of the object-oriented microarray data
analysis framework provided by the widely used marray
package in BioConductor. It is also written in R. hoptag
and its auxiliary data package, hoptagInfo, are available for
download
under version 2 of the
GNU Public License.
marray, hoptagInfo, and hoptag should be
installed in that order. Examples of usage are provided in source
code that can also be downloaded.
Q2. I am not up to installing software and learning a programming language. What can I do? A2. The best way is to collaborate with a local
statistician or bioinformatics specialist. Although it is currently
not practical for us to
provide a comprehensive tech support service, please feel free to
contact us by email if questions do arise so that we can try to
help you. The feedback will also help us identify those issues that
need the most attention.
Q3. Can you describe the microarray design in more detail? A3. Briefly, all UpTags and DnTags (upstream and downstream
tags) available in Yeast Knockout strains are represented on a
single 22,575 feature glass slide. In addition to these ~12,000
features, the UpTags and DnTags for 800 genes are replicated five
times. UpTags and DnTags for 159 "YQLnnnC" (fictitious) genes are
also replicated five times as independent negative controls. The
array design (a 1.4 MB file) is available from the Gene Expression
Omnibus (GEO)
microarray data repository at
NCBI
(accession number
GPL1444).
For detailed information about the design and validation of the
array, please see our recent paper in
Nucleic Acids Research.
Q4. What annotation files are available for analyzing data from microarray scanners? A4. We currently offer the following types of annotation files,
each provided in 'zip' (compressed) format. Please let us know if
these present any difficulties.
Agilent scanners:
GEML (0.7 MB) |
MAGE-ML (2.9 MB)
GenePix scanners:
GAL (0.25 MB)
Q5. What information is available about poorly hybridizing TAGs in the heterozgous YKO strain collection? A5.We have surveyed TAG behavior from 1121 hybridizations.
Some of the results from this study are shown here. Cy3 and Cy5
intensity values were quantile normalized
(Bolstad et al., 2003),
and the median absolute deviation (MAD, a robust estimate of the
standard deviation) for each TAG across all scans and colors was
plotted versus mean log2 intensity (see
UPTAG
and DNTAG plots).
TAGs with MAD below the indicated line are annotated as "failed"; the
list can be downloaded
here."
This definition is such that failed TAGs are consistent
(never change). TAGs that never change also coincide with low signal
TAGs, reinforcing the view that these TAGs are non-functional.
Histograms
of the MADs for UPTAGs and DNTAGs are shown, together
with the cutoff values defining failed TAGs (0.40 and 0.35 for
UPTAGs and DNTAGs, respectively).
Q6. How was this website put together? A6. Much Open Source technology is at work -- a
PostgreSQL database, the
DarwinPorts and
Fink software repositories
for Mac OS X, the
Python programming language
in conjunction with the
PyPgSQL
PostgreSQL adapter, and the
HTMLgen
HTML generator. Many thanks to these developers (and others, not to
mention
R) for the
wonderful tools.
The gene names used here conform to those at the Saccharomyces
Genome Database
(SGD),
as downloaded 04/18/05.
Q1. In addition to the posted candidate lists on the search page, will raw data be available? A1. We are planning to post these data to
GEO and will reference
those links
here.
