bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Дек. 31, 2024
Abstract
mRNA
delivery
offers
new
opportunities
for
disease
treatment
by
directing
cells
to
produce
therapeutic
proteins.
However,
designing
highly
stable
mRNAs
with
programmable
cell
type-specificity
remains
a
challenge.
To
address
this,
we
measured
the
regulatory
activity
of
60,000
5’
and
3’
untranslated
regions
(UTRs)
across
six
types
developed
PARADE
(Prediction
And
RAtional
DEsign
UTRs),
generative
AI
framework
engineer
RNA
tailored
type-specific
activity.
We
validated
testing
15,800
de
novo-designed
sequences
these
lines
identified
many
that
demonstrated
superior
specificity
compared
existing
therapeutics.
PARADE-engineered
UTRs
also
exhibited
robust
tissue-specific
in
animal
models,
achieving
selective
expression
liver
spleen.
leveraged
enhance
stability,
significantly
increasing
protein
output
durability
vivo.
These
advancements
translated
notable
increases
efficacy,
as
PARADE-designed
oncosuppressor
mRNAs,
namely
PTEN
P16,
effectively
reduced
tumor
growth
patient-derived
neuroglioma
xenograft
models
orthotopic
mouse
models.
Collectively,
findings
establish
versatile
platform
safer,
more
precise,
therapies.
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Ноя. 15, 2024
Deep
neural
networks
(DNNs)
have
advanced
predictive
modeling
for
regulatory
genomics,
but
challenges
remain
in
ensuring
the
reliability
of
their
predictions
and
understanding
key
factors
behind
decision
making.
Here
we
introduce
DEGU
(Distilling
Ensembles
Genomic
Uncertainty-aware
models),
a
method
that
integrates
ensemble
learning
knowledge
distillation
to
improve
robustness
explainability
DNN
predictions.
distills
an
DNNs
into
single
model,
capturing
both
average
ensemble's
variability
across
them,
with
latter
representing
epistemic
(or
model-based)
uncertainty.
also
includes
optional
auxiliary
task
estimate
aleatoric,
or
data-based,
uncertainty
by
experimental
replicates.
By
applying
various
functional
genomic
prediction
tasks,
demonstrate
DEGU-trained
models
inherit
performance
benefits
ensembles
improved
generalization
out-of-distribution
sequences
more
consistent
explanations
cis-regulatory
mechanisms
through
attribution
analysis.
Moreover,
provide
calibrated
estimates,
conformal
offering
coverage
guarantees
under
minimal
assumptions.
Overall,
paves
way
robust
trustworthy
applications
deep
genomics
research.
Cells,
Год журнала:
2024,
Номер
13(23), С. 1963 - 1963
Опубликована: Ноя. 27, 2024
Gene
therapy
is
a
promising
approach
to
the
treatment
of
various
inherited
diseases,
but
its
development
complicated
by
number
limitations
natural
promoters
used.
The
currently
used
strong
ubiquitous
do
not
allow
for
specificity
expression,
while
tissue-specific
have
lowactivity.
These
can
be
addressed
creating
new
synthetic
that
achieve
high
levels
target
gene
expression.
This
review
discusses
recent
advances
in
provide
more
precise
regulation
Approaches
design
are
reviewed,
including
manual
and
bioinformatic
methods
using
machine
learning.
Examples
successful
applications
hereditary
diseases
cancer
presented,
as
well
prospects
their
clinical
use.
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Дек. 11, 2024
Precise
control
of
gene
expression
is
essential
for
cellular
function,
but
the
mechanisms
by
which
enhancers
communicate
with
promoters
to
coordinate
this
process
are
not
fully
understood.
While
sequence-based
deep
learning
models
show
promise
in
predicting
enhancer-driven
expression,
experimental
validation
and
human-interpretable
mechanistic
insights
lag
behind.
Here,
we
present
EXTRA-seq
,
a
novel
EXT
ended
R
eporter
A
ssay
followed
seq
uencing
designed
quantify
enhancer
activity
endogenous
contexts
over
kilobase-scale
distances.
We
demonstrate
that
can
be
targeted
disease-relevant
loci
captures
changes
at
resolution
individual
transcription
factor
binding
sites,
enabling
discovery.
Using
engineered
synthetic
enhancer-promoter
combinations,
reveal
TATA-box
acts
as
dynamic
range
amplifier,
modulating
levels
function
strength.
Importantly,
find
integrating
state-of-the-art
plasmid-based
assays
improves
prediction
measured
EXTRA-seq.
These
findings
open
new
avenues
predictive
modeling
therapeutic
applications.
Overall,
our
work
provides
powerful
platform
interrogate
complex
interplay
between
promoters,
bridging
gap
silico
predictions
biological
mechanisms.
bioRxiv (Cold Spring Harbor Laboratory),
Год журнала:
2024,
Номер
unknown
Опубликована: Дек. 31, 2024
Abstract
mRNA
delivery
offers
new
opportunities
for
disease
treatment
by
directing
cells
to
produce
therapeutic
proteins.
However,
designing
highly
stable
mRNAs
with
programmable
cell
type-specificity
remains
a
challenge.
To
address
this,
we
measured
the
regulatory
activity
of
60,000
5’
and
3’
untranslated
regions
(UTRs)
across
six
types
developed
PARADE
(Prediction
And
RAtional
DEsign
UTRs),
generative
AI
framework
engineer
RNA
tailored
type-specific
activity.
We
validated
testing
15,800
de
novo-designed
sequences
these
lines
identified
many
that
demonstrated
superior
specificity
compared
existing
therapeutics.
PARADE-engineered
UTRs
also
exhibited
robust
tissue-specific
in
animal
models,
achieving
selective
expression
liver
spleen.
leveraged
enhance
stability,
significantly
increasing
protein
output
durability
vivo.
These
advancements
translated
notable
increases
efficacy,
as
PARADE-designed
oncosuppressor
mRNAs,
namely
PTEN
P16,
effectively
reduced
tumor
growth
patient-derived
neuroglioma
xenograft
models
orthotopic
mouse
models.
Collectively,
findings
establish
versatile
platform
safer,
more
precise,
therapies.
