Nature Communications,
Год журнала:
2022,
Номер
13(1)
Опубликована: Дек. 1, 2022
Abstract
Single-cell
data
integration
can
provide
a
comprehensive
molecular
view
of
cells.
However,
how
to
integrate
heterogeneous
single-cell
multi-omics
as
well
spatially
resolved
transcriptomic
remains
major
challenge.
Here
we
introduce
uniPort,
unified
framework
that
combines
coupled
variational
autoencoder
(coupled-VAE)
and
minibatch
unbalanced
optimal
transport
(Minibatch-UOT).
It
leverages
both
highly
variable
common
dataset-specific
genes
for
handle
the
heterogeneity
across
datasets,
it
is
scalable
large-scale
datasets.
uniPort
jointly
embeds
datasets
into
shared
latent
space.
further
construct
reference
atlas
gene
imputation
Meanwhile,
provides
flexible
label
transfer
deconvolute
spatial
using
an
plan,
instead
embedding
We
demonstrate
capability
by
applying
variety
including
transcriptomics,
chromatin
accessibility,
data.
Nature,
Год журнала:
2023,
Номер
619(7970), С. 572 - 584
Опубликована: Июль 19, 2023
Abstract
The
intestine
is
a
complex
organ
that
promotes
digestion,
extracts
nutrients,
participates
in
immune
surveillance,
maintains
critical
symbiotic
relationships
with
microbiota
and
affects
overall
health
1
.
intesting
has
length
of
over
nine
metres,
along
which
there
are
differences
structure
function
2
localization
individual
cell
types,
type
development
trajectories
detailed
transcriptional
programs
probably
drive
these
function.
Here,
to
better
understand
differences,
we
evaluated
the
organization
single
cells
using
multiplexed
imaging
single-nucleus
RNA
open
chromatin
assays
across
eight
different
intestinal
sites
from
donors.
Through
systematic
analyses,
find
compositions
differ
substantially
regions
demonstrate
complexity
epithelial
subtypes,
same
types
organized
into
distinct
neighbourhoods
communities,
highlighting
immunological
niches
present
intestine.
We
also
map
gene
regulatory
suggestive
differentiation
cascade,
associate
disease
heritability
specific
types.
These
results
describe
composition,
regulation
for
this
organ,
serve
as
an
important
reference
understanding
human
biology
disease.
Cell,
Год журнала:
2023,
Номер
186(20), С. 4422 - 4437.e21
Опубликована: Сен. 1, 2023
Recent
work
has
identified
dozens
of
non-coding
loci
for
Alzheimer's
disease
(AD)
risk,
but
their
mechanisms
and
AD
transcriptional
regulatory
circuitry
are
poorly
understood.
Here,
we
profile
epigenomic
transcriptomic
landscapes
850,000
nuclei
from
prefrontal
cortexes
92
individuals
with
without
to
build
a
map
the
brain
regulome,
including
profiles,
regulators,
co-accessibility
modules,
peak-to-gene
links
in
cell-type-specific
manner.
We
develop
methods
multimodal
integration
detecting
modules
using
linking.
show
risk
enriched
microglial
enhancers
specific
TFs
SPI1,
ELF2,
RUNX1.
detect
9,628
ATAC-QTL
loci,
which
integrate
alongside
prioritize
variant
circuits.
report
differential
accessibility
late
glia
early
neurons.
Strikingly,
late-stage
brains
global
epigenome
dysregulation
indicative
erosion
cell
identity
loss.
Nature Communications,
Год журнала:
2023,
Номер
14(1)
Опубликована: Фев. 21, 2023
Abstract
Single-cell
multi-omics
(scMulti-omics)
allows
the
quantification
of
multiple
modalities
simultaneously
to
capture
intricacy
complex
molecular
mechanisms
and
cellular
heterogeneity.
Existing
tools
cannot
effectively
infer
active
biological
networks
in
diverse
cell
types
response
these
external
stimuli.
Here
we
present
DeepMAPS
for
network
inference
from
scMulti-omics.
It
models
scMulti-omics
a
heterogeneous
graph
learns
relations
among
cells
genes
within
both
local
global
contexts
robust
manner
using
multi-head
transformer.
Benchmarking
results
indicate
performs
better
than
existing
clustering
construction.
also
showcases
competitive
capability
deriving
cell-type-specific
lung
tumor
leukocyte
CITE-seq
data
matched
diffuse
small
lymphocytic
lymphoma
scRNA-seq
scATAC-seq
data.
In
addition,
deploy
webserver
equipped
with
functionalities
visualizations
improve
usability
reproducibility
analysis.
Nature Biotechnology,
Год журнала:
2023,
Номер
41(3), С. 399 - 408
Опубликована: Янв. 2, 2023
The
application
of
multiple
omics
technologies
in
biomedical
cohorts
has
the
potential
to
reveal
patient-level
disease
characteristics
and
individualized
response
treatment.
However,
scale
heterogeneous
nature
multi-modal
data
makes
integration
inference
a
non-trivial
task.
We
developed
deep-learning-based
framework,
multi-omics
variational
autoencoders
(MOVE),
integrate
such
applied
it
cohort
789
people
with
newly
diagnosed
type
2
diabetes
deep
phenotyping
from
DIRECT
consortium.
Using
silico
perturbations,
we
identified
drug-omics
associations
across
datasets
for
20
most
prevalent
drugs
given
substantially
higher
sensitivity
than
univariate
statistical
tests.
From
these,
among
others,
novel
between
metformin
gut
microbiota
as
well
opposite
molecular
responses
two
statins,
simvastatin
atorvastatin.
used
quantify
drug-drug
similarities,
assess
degree
polypharmacy
conclude
that
drug
effects
are
distributed
modalities.
