Please wait while the transcript is being prepared...
0:00
Hello and welcome
to this talk on
AI in Drug Discovery.
My name is Ed Addison.
I'm the chairman of
Cloud Pharmaceuticals.
I also advise several AI and
drug discovery companies,
and I lecture at NC
State University.
0:19
Today I'm going to talk about
methods of AI in drug discovery.
Primarily, I'm going to focus on
what the methods are
and where they apply.
I'm not going to
go into details on
algorithms or at a
technical level.
That I assume that
you have access to
do that on your own.
Here instead, I'm going
to talk primarily about
the philosophy of AI
in drug discovery and
where the benefits are.
You're going to primarily
hear about four benefits.
One being that AI makes drug
discovery more efficient,
in other words, faster
and lower cost.
You're also going to hear that
AI reduces the failure rate,
that's even more important
because it acts as
a lever arm on the total
cost of drug development.
Lastly, AI can
generate novel IP.
1:13
By AI, I'm going to cover both
general AI machine learning
and also generative AI.
But as you know, generative
AI is a major force today.
I want to point out that
generative AI is
predicted to have
a $160 billion impact on
the pharmaceutical
industry as stated by
McKinsey in a recent report.
1:40
I'm going to start by
talking about small
molecule design.
There are many places
where AI can apply in
the drug discovery and
development pipeline,
from discovery of targets
to leads to biomarkers to
preclinical data analysis through
applications in the clinic.
I'm going to identify
methods from all of these.
It's not intended to be
comprehensive of everything,
but rather to give you
a sampler of methods that apply,
to give you a general
idea of what applies and
where the benefits are.
In small molecule design,
we can use machine learning
or generative AI or
generative adversarial
networks to
accelerate molecular screening.
In this case, AI
is often used as
a software implementation of
high-throughput screening,
which is a cost-saving approach.
It can also be used to
optimize molecular properties.
This assists during the
lead optimization phase.
It can be used in drug design.
In other words,
tailoring drug-like
characteristics so that
you can come up with
novel IP for the small
molecule design.
