Chemical Reviews,
Год журнала:
2021,
Номер
122(2), С. 2907 - 2980
Опубликована: Сен. 24, 2021
In
the
pursuit
of
new
pharmaceuticals
and
agrochemicals,
chemists
in
life
science
industry
require
access
to
mild
robust
synthetic
methodologies
systematically
modify
chemical
structures,
explore
novel
space,
enable
efficient
synthesis.
this
context,
photocatalysis
has
emerged
as
a
powerful
technology
for
synthesis
complex
often
highly
functionalized
molecules.
This
Review
aims
summarize
published
contributions
field
from
industry,
including
research
industrial-academic
partnerships.
An
overview
developed
strategic
applications
synthesis,
peptide
functionalization,
isotope
labeling,
both
DNA-encoded
traditional
library
is
provided,
along
with
summary
state-of-the-art
photoreactor
effective
upscaling
photocatalytic
reactions.
Chemical Reviews,
Год журнала:
2017,
Номер
117(18), С. 11796 - 11893
Опубликована: Июнь 1, 2017
Flow
chemistry
involves
the
use
of
channels
or
tubing
to
conduct
a
reaction
in
continuous
stream
rather
than
flask.
equipment
provides
chemists
with
unique
control
over
parameters
enhancing
reactivity
some
cases
enabling
new
reactions.
This
relatively
young
technology
has
received
remarkable
amount
attention
past
decade
many
reports
on
what
can
be
done
flow.
Until
recently,
however,
question,
"Should
we
do
this
flow?"
merely
been
an
afterthought.
review
introduces
readers
basic
principles
and
fundamentals
flow
critically
discusses
recent
accounts.
Chemical Society Reviews,
Год журнала:
2015,
Номер
45(3), С. 546 - 576
Опубликована: Окт. 28, 2015
The
advent
of
modern
C-H
functionalization
chemistries
has
enabled
medicinal
chemists
to
consider
a
synthetic
strategy,
late
stage
(LSF),
which
utilizes
the
bonds
drug
leads
as
points
diversification
for
generating
new
analogs.
LSF
approaches
offer
promise
rapid
exploration
structure
activity
relationships
(SAR),
generation
oxidized
metabolites,
blocking
metabolic
hot
spots
and
preparation
biological
probes.
This
review
details
toolbox
intermolecular
with
proven
applicability
drug-like
molecules,
classified
by
regioselectivity
patterns,
gives
guidance
on
how
systematically
develop
strategies
using
these
patterns
other
considerations.
In
addition,
number
examples
illustrate
have
been
used
impact
actual
discovery
chemical
biology
efforts.
Science,
Год журнала:
2018,
Номер
360(6385), С. 186 - 190
Опубликована: Фев. 15, 2018
A
guide
for
catalyst
choice
in
the
forest
Chemists
often
discover
reactions
by
applying
catalysts
to
a
series
of
simple
compounds.
Tweaking
those
tolerate
more
structural
complexity
pharmaceutical
research
is
time-consuming.
Ahneman
et
al.
report
that
machine
learning
can
help.
Using
high-throughput
data
set,
they
trained
random
algorithm
predict
which
specific
palladium
would
best
isoxazoles
(cyclic
structures
with
an
N–O
bond)
during
C–N
bond
formation.
The
predictions
also
helped
analysis
inhibition
mechanism.
Science
,
this
issue
p.
186
Pairing
prediction
and
robotic
synthesis
Progress
in
automated
of
organic
compounds
has
been
proceeding
along
parallel
tracks.
One
goal
is
algorithmic
viable
routes
to
a
desired
compound;
the
other
implementation
known
reaction
sequence
on
platform
that
needs
little
no
human
intervention.
Coley
et
al.
now
report
preliminary
integration
these
two
protocols.
They
paired
retrosynthesis
algorithm
with
robotically
reconfigurable
flow
apparatus.
Human
intervention
was
still
required
supplement
predictor
practical
considerations
such
as
solvent
choice
precise
stoichiometry,
although
predictions
should
improve
accessible
data
accumulate
for
training.
Science
,
this
issue
p.
eaax1566
Accounts of Chemical Research,
Год журнала:
2018,
Номер
51(5), С. 1281 - 1289
Опубликована: Май 1, 2018
ConspectusComputer-aided
synthesis
planning
(CASP)
is
focused
on
the
goal
of
accelerating
process
by
which
chemists
decide
how
to
synthesize
small
molecule
compounds.
The
ideal
CASP
program
would
take
a
molecular
structure
as
input
and
output
sorted
list
detailed
reaction
schemes
that
each
connect
target
purchasable
starting
materials
via
series
chemically
feasible
steps.
Early
work
in
this
field
relied
expert-crafted
rules
heuristics
describe
possible
retrosynthetic
disconnections
selectivity
but
suffered
from
incompleteness,
infeasible
suggestions,
human
bias.
With
relatively
recent
availability
large
corpora
(such
United
States
Patent
Trademark
Office
(USPTO),
Reaxys,
SciFinder
databases),
consisting
millions
tabulated
examples,
it
now
construct
validate
purely
data-driven
approaches
planning.
As
result,
has
been
opened
machine
learning
techniques,
advancing
rapidly.In
Account,
we
focus
two
critical
aspects
both
challenges.
First,
discuss
problem
planning,
requires
recommender
system
propose
synthetic
molecule.
We
search
strategy,
necessary
overcome
exponential
growth
space
with
increasing
number
steps,
can
be
assisted
through
learned
complexity
metric.
also
recursive
expansion
performed
straightforward
nearest
neighbor
model
makes
clever
use
data
generate
high
quality
disconnections.
Second,
anticipating
products
chemical
reactions,
used
proposed
reactions
computer-generated
plan
(i.e.,
reduce
false
positives)
increase
likelihood
experimental
success.
While
introduce
task
context
validation,
its
utility
extends
prediction
side
impurities,
among
other
applications.
neural
network-based
others
have
developed
for
forward
trained
previously
published
data.Machine
artificial
intelligence
revolutionized
disciplines,
not
limited
image
recognition,
dictation,
translation,
content
recommendation,
advertising,
autonomous
driving.
there
rich
history
using
structure–activity
models
chemistry,
only
being
successfully
applied
more
broadly
organic
design.
reported
rapidly
transforming
CASP,
are
several
remaining
challenges
opportunities,
many
pertaining
standardization
evaluation
metrics,
must
addressed
community
at
large.
