Frontiers in Genetics,
Journal Year:
2022,
Volume and Issue:
13
Published: May 18, 2022
Genomic
prediction
tools
support
crop
breeding
based
on
statistical
methods,
such
as
the
genomic
best
linear
unbiased
(GBLUP).
However,
these
are
not
designed
to
capture
non-linear
relationships
within
multi-dimensional
datasets,
or
deal
with
high
dimension
datasets
imagery
collected
by
unmanned
aerial
vehicles.
Machine
learning
(ML)
algorithms
have
potential
surpass
accuracy
of
current
used
for
genotype
phenotype
prediction,
due
their
capacity
autonomously
extract
data
features
and
represent
at
multiple
levels
abstraction.
This
review
addresses
challenges
applying
machine
methods
predicting
phenotypic
traits
genetic
markers,
environment
data,
breeding.
We
present
advantages
disadvantages
explainable
model
structures,
discuss
models
in
breeding,
challenges,
including
scarcity
high-quality
inconsistent
metadata
annotation
requirements
ML
models.
The Plant Cell,
Journal Year:
2022,
Volume and Issue:
35(1), P. 162 - 186
Published: Nov. 12, 2022
Abstract
Breeding
climate-resilient
crops
with
improved
levels
of
abiotic
and
biotic
stress
resistance
as
a
response
to
climate
change
presents
both
opportunities
challenges.
Applying
the
framework
“breeder’s
equation,”
which
is
used
predict
selection
for
breeding
program
cycle,
we
review
methodologies
strategies
that
have
been
successfully
breed
drought
resistance,
where
target
population
environments
(TPEs)
spatially
temporally
heterogeneous
mixture
drought-affected
favorable
(water-sufficient)
environments.
Long-term
improvement
temperate
maize
US
corn
belt
case
study
compared
progress
other
geographies.
Integration
trait
information
across
scales,
from
genomes
ecosystems,
needed
accurately
yield
outcomes
genotypes
within
current
future
TPEs.
This
will
require
transdisciplinary
teams
explore,
identify,
exploit
novel
accelerate
outcomes;
germplasm
resources
products
(cultivars,
hybrids,
clones,
populations)
outperform
replace
in
use
by
farmers,
combination
modified
agronomic
management
suited
their
local
Heredity,
Journal Year:
2020,
Volume and Issue:
126(1), P. 92 - 106
Published: Aug. 27, 2020
Abstract
Modern
whole-genome
prediction
(WGP)
frameworks
that
focus
on
multi-environment
trials
(MET)
integrate
large-scale
genomics,
phenomics,
and
envirotyping
data.
However,
the
more
complex
statistical
model,
longer
computational
processing
times,
which
do
not
always
result
in
accuracy
gains.
We
investigated
use
of
new
kernel
methods
modeling
structures
involving
genomics
nongenomic
sources
variation
two
MET
maize
data
sets.
Five
WGP
models
were
considered,
advancing
complexity
from
a
main-effect
additive
model
(A)
to
structures,
including
dominance
deviations
(D),
genotype
×
environment
interaction
(AE
DE),
reaction-norm
using
environmental
covariables
(W)
their
with
A
D
(AW
+
DW).
combination
those
built
three
different
methods,
Gaussian
(GK),
Deep
(DK),
benchmark
genomic
best
linear-unbiased
predictor
(GBLUP/GB),
was
tested
under
scenarios:
newly
developed
hybrids
(CV1),
sparse
conditions
(CV2),
environments
(CV0).
GK
DK
outperformed
GB
reduction
computation
time
(~up
20%)
all
model–kernel
scenarios.
efficient
capturing
due
AE
DE
effects
translated
it
into
gains
85%
compared
GB).
provided
consistent
predictions,
even
for
such
as
W
AW
DW.
Our
results
suggest
are
translating
accuracy,
suitable
biologically
accurate
faster
way.
Horticulture Research,
Journal Year:
2019,
Volume and Issue:
6(1)
Published: April 5, 2019
In
2010,
a
major
scientific
milestone
was
achieved
for
tree
fruit
crops:
publication
of
the
first
draft
whole
genome
sequence
(WGS)
apple
(Malus
domestica).
This
WGS,
v1.0,
valuable
as
initial
reference
information,
fine
mapping,
gene
discovery,
variant
and
tool
development.
A
new,
high
quality
GDDH13
v1.1,
released
in
2017
now
serves
apple.
Over
past
decade,
these
WGSs
have
had
an
enormous
impact
on
our
understanding
biological
functioning,
trait
physiology
inheritance,
leading
to
practical
applications
improving
this
highly
valued
crop.
Causal
identities
phenotypes
fundamental
interest
can
today
be
discovered
much
more
rapidly.
Genome-wide
polymorphisms
at
genetic
resolution
are
screened
efficiently
over
hundreds
thousands
individuals
with
new
insights
into
relationships
pedigrees.
High-density
maps
constructed
quantitative
loci
traits
readily
associated
positional
candidate
genes
and/or
converted
diagnostic
tests
breeders.
We
understand
species,
geographical,
genomic
origins
domesticated
precisely,
well
its
relationship
wild
relatives.
The
WGS
has
turbo-charged
application
classical
research
steps
crop
improvement
drives
innovative
methods
achieve
durable,
environmentally
sound,
productive,
consumer-desirable
production.
review
includes
examples
basic
breakthroughs
challenges
using
WGSs.
Recommendations
"what's
next"
focus
necessary
upgrades
data
pool,
use
data,
reach
frontiers
genomics-based
Plant Communications,
Journal Year:
2021,
Volume and Issue:
2(6), P. 100230 - 100230
Published: Aug. 9, 2021
Genotyping
platforms,
as
critical
supports
for
genomics,
genetics,
and
molecular
breeding,
have
been
well
implemented
at
national
institutions/universities
in
developed
countries
multinational
seed
companies
that
possess
high-throughput,
automatic,
large-scale,
shared
facilities.
In
this
study,
we
integrated
an
improved
genotyping
by
target
sequencing
(GBTS)
system
with
capture-in-solution
(liquid
chip)
technology
to
develop
a
multiple
single-nucleotide
polymorphism
(mSNP)
approach
which
mSNPs
can
be
captured
from
single
amplicon.
From
one
40K
maize
mSNP
panel,
three
types
of
markers
(40K
mSNPs,
251K
SNPs,
690K
haplotypes),
generated
panels
various
marker
densities
(1K-40K
mSNPs)
different
depths.
Comparative
genetic
diversity
analysis
was
performed
genic
versus
intergenic
di-allelic
SNPs
non-typical
SNPs.
Compared
the
one-amplicon-one-SNP
system,
within-mSNP
haplotypes
are
more
powerful
detection,
linkage
disequilibrium
decay
analysis,
genome-wide
association
studies.
The
technologies,
protocols,
application
scenarios
study
will
serve
model
development
arrays
highly
efficient
GBTS
systems
animals,
plants,
microorganisms.
in silico Plants,
Journal Year:
2019,
Volume and Issue:
1(1)
Published: Jan. 1, 2019
Abstract
The
potential
to
add
significant
value
the
rapid
advances
in
plant
breeding
technologies
associated
with
statistical
whole-genome
prediction
methods
is
a
new
frontier
for
crop
physiology
and
modelling.
Yield
advance
by
genetic
improvement
continues
require
of
phenotype
based
on
genotype,
this
remains
challenging
complex
traits
despite
recent
genotyping
phenotyping.
Crop
models
that
capture
physiological
knowledge
can
robustly
predict
phenotypic
consequences
genotype-by-environment-by-management
(G×E×M)
interactions
have
demonstrated
as
an
integrating
tool.
But
does
biological
reality
come
degree
complexity
restricts
applicability
improvement?
Simple,
high-speed,
parsimonious
are
required
dealing
thousands
genotypes
environment
combinations
modern
programs
utilizing
genomic
technologies.
In
contrast,
it
often
considered
greater
model
needed
evaluate
putative
variation
specific
target
environments
their
underpinning
biology
advances.
Is
contradiction
leading
divergent
futures?
Here
argued
parsimony
do
not
need
be
independent
perhaps
should
be.
Models
structured
readily
allow
level
process
algorithms,
while
using
coding
computational
facilitate
high-speed
simulation,
could
well
provide
structure
next
generation
support
enhance
Beyond
that,
trans-scale
transdisciplinary
dialogue
among
scientists
will
construct
such
effectively
at
least
important
models.
The Crop Journal,
Journal Year:
2021,
Volume and Issue:
9(3), P. 669 - 677
Published: April 22, 2021
Rice
(Oryza
sativa)
provides
a
staple
food
source
for
more
than
half
the
world
population.
However,
current
pace
of
rice
breeding
in
yield
growth
is
insufficient
to
meet
demand
ever-increasing
global
Genomic
selection
(GS)
holds
great
potential
accelerate
progress
and
cost-effective
via
early
before
phenotypes
are
measured.
Previous
simulation
experimental
studies
have
demonstrated
usefulness
GS
breeding.
several
affecting
factors
limitations
require
careful
consideration
when
performing
GS.
In
this
review,
we
summarize
major
genetics
statistical
predictive
performance
as
well
application
We
also
highlight
effective
strategies
increase
ability
various
models,
including
models
incorporating
functional
markers,
genotype
by
environment
interactions,
multiple
traits,
index,
omic
data.
Finally,
envision
that
integrating
with
other
advanced
technologies
such
unmanned
aerial
vehicles
open-source
platforms
will
further
improve
efficiency
reduce
cost
Agronomy,
Journal Year:
2020,
Volume and Issue:
10(4), P. 585 - 585
Published: April 19, 2020
Sugarcane
is
a
major
industrial
crop
cultivated
in
tropical
and
subtropical
regions
of
the
world.
It
primary
source
sugar
worldwide,
accounting
for
more
than
70%
world
consumption.
Additionally,
sugarcane
emerging
as
sustainable
bioenergy.
However,
increase
productivity
from
has
been
small
compared
to
other
crops,
rate
genetic
gains
current
breeding
programs
tends
be
plateauing.
In
this
review,
some
main
contributors
relatively
slow
rates
gain
are
discussed,
including
(i)
cycle
length
(ii)
low
narrow-sense
heritability
commercial
traits,
possibly
reflecting
strong
non-additive
effects
involved
quantitative
trait
expression.
A
general
overview
genomic
selection
(GS),
modern
tool
that
very
successfully
applied
animal
plant
breeding,
given.
This
review
discusses
key
elements
GS
its
potential
significantly
sugarcane,
mainly
by
reducing
length,
increasing
prediction
accuracy
clonal
performance,
(iii)
values
parent
selection.
approaches
can
accurately
capture
potentially
improve
estimated
particularly
promising
adoption
breeding.
Finally,
different
strategies
efficient
incorporation
practical
context
presented.
These
proposed
hold
substantially
future
