How Big of an issue is it?
My name is Peter Welch and I'm an R&D director at Invitrogen.
There are many things to consider when looking at siRNA specificity.
I'll cover five different topics,
the first of which is
siRNA design and activity.
And secondly, I'll talk about ways of
inactivating the sense-strand of the siRNA molecule,
to essentially remove half of those potential off-target molecules.
Third, I'll talk about avoiding the interferon stress-response in mammalian cells.
Number four, I'll talk about new analysis called
Smith-Waterman which can remove a lot of the cross-reactive genes.
And then, finally, I'll end with talking about proper experimental design,
and make some suggestions about different controls to include,
and ways to avoid and control other specificity questions
that might be in your RNAi experiments.
But first, let me just briefly review using RNAi in mammalian cells.
Many of you I'm sure are familiar with the use of standard siRNAs.
These are typically 21-mers or 19-mers with overhangs,
and these are double-stranded RNA molecules that are
transfected into a mammalian cell or delivered.
And then, that double-stranded RNA is unwound by a multi-protein complex called RISC,
which stands for RNA-Induced Silencing Complex.
Now, what the RISC complex does is that as it comes to that duplex,
it unwinds it and it captures the antisense strand which is shown in red.
That antisense strand has complementarity or
sequence complementarity to the messenger RNA you would like to target for destruction.
And that, ultimately leads to the cleavage of the messenger RNA,
destabilizing it, and it's destroyed in the cell,
and that's how you get your gene knockdown.
There are other ways of using RNAi in mammalian cells.
Some people have recently found that longer than 21-mers,
actually have either increased activity or other benefits.
And, I will talk a little bit about a technology called Stealth RNAi,
that's actually 25-mer blunt molecules but they enter the exact same pathway.
Now, what I think was surprising when it comes to specificity is that,
these 21 or 25-mer sequences,
statistically, should be unique in the whole human genome.
And so, one would expect at first glance that these are going to be extremely specific,
all through Watson-Crick base pairing,
and they won't be targeting any other genes in the cell,
if you design them properly.
Unfortunately, that's turned out not to be
the case and that's what I'll describe in the rest of the talk,
and talk about ways that you can prevent that.
What I'm not going to touch on are the other ways of using RNAi in a mammalian cell,
either by short hairpin or shRNAs,
micro RNAs or even the processing of long double-stranded RNA by
dicer in vitro and then transfecting these pools of diced molecules.
But the design rules are very similar for these other molecules.
To begin with, I'd like to set just a few definitions so