Prediction of Respiratory Irritation and Respiratory Sensitization of Chemicals Using Structural Alerts and Machine Learning Modeling
Toxics,
Год журнала:
2025,
Номер
13(4), С. 243 - 243
Опубликована: Март 25, 2025
Inhalation
of
toxic
substances
and
contaminants
can
have
adverse
effects
on
the
respiratory
tract,
leading
to
a
range
health
problems,
such
as
irritation
inflammation,
allergic
reaction
asthma,
lung
damage,
or
even
death.
It
is
not
possible
experimentally
evaluate
toxicity
for
all
thousands
chemicals
in
use.
Here,
we
generated
structural
alerts
developed
machine
learning
(ML)
classification
models
predict
sensitization
hazards
using
experimental
data
from
publicly
available
databases
literature.
We
identified
13
irritants
18
sensitizers.
also
set
each
hazard
different
types
molecular
descriptors
ML
techniques.
Five
best
performing
were
combined
into
consensus
model
irritation,
four
individual
used
develop
sensitization.
The
prediction
accuracy
was
84%
training
88%
test
set,
86%
both
sets.
A
combination
screen
occupational-
military-relevant
chemicals.
Out
687
screened
occupational
chemicals,
62
compounds
121
sensitizers,
while
47
predicted
36
sensitizers
list
525
Язык: Английский
Endpoints Lacking Animal-Free Alternatives Under REACH and the Remaining Challenges
Deleted Journal,
Год журнала:
2025,
Номер
2(2)
Опубликована: Янв. 1, 2025
The
European
Union’s
REACH
regulation
mandates
comprehensive
chemical
testing
to
ensure
human
and
environmental
safety,
prioritizing
non-animal
methods
whenever
possible.
Despite
significant
advancements
in
alternative
strategies,
certain
endpoints
still
require
animal
studies
due
the
lack
of
validated
vitro,
ex
vivo
or
silico
models.
This
paper
examines
these
gaps
by
tonnage
band,
highlighting
key
areas
where
alternatives
remain
insufficient.
Acute
oral,
dermal,
inhalation
toxicity
tests,
repeated-dose
as
well
reproductive
developmental
are
among
that
reliant
on
testing.
While
progress
has
been
made
developing
such
Defined
Approaches
(DAs)
may
integrate
physiologically
based
kinetic
(PBK)
modeling,
not
yet
fully
for
regulatory
acceptance.
Continued
investment
research,
validation
international
collaboration
is
essential
accelerate
implementation
animal-free
alternatives.
Язык: Английский