Please wait while the transcript is being prepared...
0:00
Hello, my name is John Van Drie, I'm an independent consultant in the Boston area, working with biotech companies.
My company is Van Drie Research and my topic today is virtual screening.
0:14
The outline of what I'd like to talk about today is shown here.
First, I need to explain what virtual screening is,
and I will describe three different flavors of
virtual screening and give you examples of each approach.
Next, I'd like to talk about why we use virtual screening,
what are the advantages of using this over other approaches?
Point 3, the considerations when
thinking of employing virtual screening in a project.
There are certain requirements one has to meet
before you can proceed, and I'll try to help give you
a sense for what it takes to bring virtual screening into practice in a project.
Point 4 is to answer the question: does virtual screening work?
It's an amazing thing to me that we've been using this now
over 30 years, and people still keep asking this question:
does virtual screening work?
I'll give you two case studies that have both been published recently, that I
think each illustrate very nicely the fact that it works and how it works.
Point 5, I'll give you some sense for what the outlook for the future is,
and then I'll finish up with a summary and some conclusions.
1:26
What is virtual screening?
Basically this entails selecting a small number of readily available molecules,
and submitting them for biological testing through a low-throughput, high-quality assay.
The trick is choosing these molecules in a way that
suggests they're likely to be active in those tests.
This contrasts with high-throughput screening where you test, at random,
all readily-available molecules.
You just bang through lots of plates of different molecules, without
any attempt to rationalize what molecules might be active, and so forth.
This clever selection of readily-available molecules
is usually done with a computer, nowadays considering millions of possibilities.
As I mentioned, there are three different ways that one can approach this.
The first is choosing based on similarity to a known active,
if you have a known molecule active against the target of interest,
but say there are some issues with it or off-target liabilities,
you can use virtual screening to find
similar molecules that you can then test to see if they have activity.
The second approach is choosing based on a pharmacophore,
a 3D pattern of functional groups that are essential for biological activity.
The final way is the way that people often think of virtual screening,
which is choosing a molecule based on how well it
complements a pocket in a target protein whose 3D structures are already known.
Algorithmically, this is called 'docking' and it's
part of a more general set of things, virtual screening.
Just to show that graphically,
