|
Dharmacon
Announces Initial Effort to Establish RNAi Standards
January
26, 2007—SAN
FRANCISCO–Following the lead of the microarray
community, Dharmacon has announced initial results of
their efforts to develop experimental guidelines and
standards for RNA interference (RNAi) screens. The
next step, the company says, is to gather input from
other thought leaders in the field and incorporate
those ideas into the current draft.
Dharmacon made the announcement this week at
Cambridge Healthtech Institute
’s High Content Analysis meeting.
Like
the Minimum
Information About a Microarray Experiment (MIAME)
standards, the Dharmacon-led group, called RNAi
Global wants to develop guidelines that enable
researchers to compare genome-wide screen results
between experiments and laboratories. The guidelines
will be called Minimum Information About an RNAi
Experiment (MIARE) and preliminary information can be
found at www.miare.org.
“Four
or five years ago no one was doing these types of
screens,” said Thomas J. Murphy, a field application
scientist at Dharmacon who presented the work. “Now
there are many labs across the world performing
genome-wide screens. The problem is that there is no
set of common guidelines or standards by which people
can do cross-screen comparisons. It is like everyone
is speaking their own language.”
With
this in mind, Dharmacon brought together a group of
researchers who had purchased one or more of the
company’s genome-wide siRNA libraries and asked them
to help develop an initial proposal for such guidelines.
Dharmacon also asked 10 of the contributing
laboratories to participate in
an experiment to determine how much variability would
occur if the same siRNA genome-wide screen was
performed in multiple locations. To try to minimize
variability, all of the groups were given the same
reagents and protocols.
“It
was clear that even when you make it the best case
scenario – which in reality never exists – there
is still great variability, which means that if each
one of these labs were going to publish their own set
of findings, they would likely all report very
different findings,” continued Murphy. “There would be
some common hits, but they would have very different
results. If they didn’t describe in great detail,
following a set of standards, how they actually
perform their experiments you would never be able to
explain why they saw different results. I think that
really highlights why we need this set of
standards.”
The
sorts of metrics Murphy would like to see in place
include the Z-factor, which gives a measure of the
robustness of each assay. (Despite the fact that the
labs that participated in the best-case- scenario
experiment were all highly trained, Murphy conceded
that the Z-factor for
some was very low.) Additionally, the
statistical methods used to identify hits need to be
described when the data are reported. Two of the
leading methods in discussion at this point are the
Z-score, which can account for plate-to-plate
variability and is easy to calculate, and the B-score,
which can account for both inter- and intra-plate
variability but is harder to calculate. Both scores
serve as a measure of how far away from the mean the
putative hit is.
Murphy
emphasized that such statistical measures have been
used for a long time in high-throughput screening with
small molecules, but that they are only now starting
to be appreciated by people trained in biology.
Additionally,
Dharmacon would like to see the raw data deposited in
a shared repository and available to other
researchers.
The
company will present further detail on their results and strategy
for MIARE at a later date
and will solicit comments from other thought leaders
on how to refine and adopt the guidelines.
URL: http://www.pharmadd.com/topnews/January
26 2007.asp
|
