bioRxiv (Cold Spring Harbor Laboratory),
Journal Year:
2024,
Volume and Issue:
unknown
Published: Nov. 3, 2024
Abstract
Optimizing
operational
conditions
for
complex
biological
systems
used
in
life
sciences
research
and
biotechnology
is
an
arduous
task.
Here,
we
have
applied
a
Bayesian
Optimization-based
iterative
framework
experimental
design
to
accelerate
cell
culture
media
development
two
applications.
First,
show
this
approach
yields
new
compositions
of
with
cytokine
supplementation
maintain
the
viability
distribution
PBMCs
culture.
Second,
optimize
production
three
recombinant
proteins
K.phaffii
cultivations.
For
both
applications,
identified
improved
outcomes
compared
initial
standard
using
3
30
times
fewer
experiments
than
other
methods
such
as
Design
Experiments.
Subsequently,
also
demonstrated
extensibility
our
efficiently
account
additional
factors
through
transfer
learning.
These
examples
demonstrate
how
coupling
data
collection,
modeling,
optimization
paradigm,
while
exploration-exploitation
tradeoff
each
iteration,
can
reduce
time
resources
these
types
optimizations.
Biotechnology and Bioengineering,
Journal Year:
2025,
Volume and Issue:
unknown
Published: March 5, 2025
Bayesian
optimization
is
a
stochastic,
global
black-box
algorithm.
By
combining
Machine
Learning
with
decision-making,
the
algorithm
can
optimally
utilize
information
gained
during
experimentation
to
plan
further
experiments-while
balancing
exploration
and
exploitation.
Although
Design
of
Experiments
has
traditionally
been
preferred
method
for
optimizing
bioprocesses,
AI-driven
tools
have
recently
drawn
increasing
attention
within
bioprocess
engineering.
This
review
presents
principles
methodologies
focuses
on
its
application
various
stages
engineering
in
upstream
downstream
processing.
Materials Research Letters,
Journal Year:
2024,
Volume and Issue:
12(4), P. 235 - 263
Published: Feb. 21, 2024
Additive
manufacturing
(AM),
especially
Laser
Powder-Bed
Fusion
(L-PBF),
provides
alloys
with
unique
properties,
but
faces
printability
challenges
like
porosity
and
cracks.
To
address
these
issues,
a
co-design
strategy
integrates
chemistry
process
indicators
to
efficiently
screen
the
design
space
for
defect-free
combinations.
Physics-based
models
visualization
tools
explore
space,
KGT
guide
microstructural
design.
The
approach
combines
experiments,
databases,
deep
learning
models,
Bayesian
optimization
streamline
AM
alloy
co-design.
By
merging
computational
data-driven
techniques
this
integrated
addresses
drives
future
advancements.
Analytical Chemistry,
Journal Year:
2024,
Volume and Issue:
96(33), P. 13699 - 13709
Published: July 9, 2024
In
recent
decades,
there
has
been
a
growing
interest
in
fully
automated
methods
for
tackling
complex
optimization
problems
across
various
fields.
Active
learning
(AL)
and
its
variant,
assisted
active
(AAL),
incorporating
guidance
or
assistance
from
external
sources
into
the
process,
play
key
roles
this
automation
by
enabling
autonomous
selection
of
optimal
experimental
conditions
to
efficiently
explore
problem
space.
These
approaches
are
particularly
valuable
situations
wherein
experimentation
is
costly
time-consuming.
This
study
explores
application
AAL
model-based
method
development
(MD)
liquid
chromatography
(LC)
using
Bayesian
statistics
incorporate
historical
data
analyte
information
generation
initial
retention
models.
The
process
involves
updating
model
parameters
based
on
new
experiments,
coupled
with
an
choose
most
informative
experiment
run
subsequent
step.
iterative
balances
exploitation
exploration
until
satisfactory
separation
achieved.
effectiveness
approach
demonstrated
via
two
practical
examples,
resulting
optimized
separations
limited
number
experiments
optimizing
gradient
slope.
It
shown
that
ability
leverage
past
knowledge
compound
improve
accuracy
reduce
runs
offers
flexible
alternative
fixed
design
methods.
