GRAPHIC: K. HOLOSKI/
SCIENCE
BASED ON T. S. SUMIDA AND D. A. HAFLER
SCIENCE science.orgscriptomic data to better determine cell
type–specific eQTLs and highlighted the
potential of the scRNA-seq approach over
bulk RNA-seq. However, the numbers of
individuals and cells analyzed by this study
were relatively small. Therefore, a larger
population-scale scRNA-seq study with ap-
preciable cell numbers per individual was
warranted. The studies by Yazar et al. and
Perez et al. addressed this need.
Yazar et al. characterized the transcrip-
tome and genetic variation across a total
of 1,267,758 PBMCs from 982 individuals.
Their eQTL mapping at single cell resolu-
tion enabled the identification of 117 loci
outside of the major histocompatibility
complex (MHC) region that exert cell type-
specific causal effects that account for
autoimmune disease risk. Perez et al. ap-
plied the same method for patients with
the autoimmune disease systemic lupus
erythematosus (SLE), profiling 1,263,676
PBMCs from 264 individuals (162 cases, 99
controls). Of note, in the SLE dataset, Perez
et al. observed that cell type–specific eQTL
effects of SLE risk variants were highly en-
riched in classical monocytes and B cells.
Moreover, by using type I interferon (IFN-I)
response gene expression as a proxy for
IFN-I–induced activation, they found that
the cell type–specific cis-eQTL effect was
modified by IFN-I responses. Given that an
IFN-I signature has been observed in SLE
patients, these results highlight the impor-
tance of disease-relevant cellular and bio-logical contexts to better understand the
disease-associated genetic effects.
A distinct feature of the scRNA-seq–
based approach over bulk RNA-seq is the
ability to compute dynamic transcriptional
transitions of cellular state and project
cells onto an axis called pseudotime that
represents the progression trajectory. By
inferring the dynamic trajectory of cells,
the effects of eQTLs on this cellular dy-
namism can be investigated. Yazar et al.
applied this method to identify dynamic
eQTLs during B cell maturation that were
not detected by cell type–specific cis-eQTL
analysis. In addition, RNA velocity, which
allows inference of future transcriptional
state direction, can be integrated with ge-
netic effects in scRNA-seq data. These new
features, which can only be achieved by
scRNA-seq–based approaches, could ex-
pand our understanding of context-depen-
dent genetic effects beyond the framework
of conventional cell type–specific effects
(see the figure).
Although scRNA-seq has advantages,
several shortcomings are apparent. The
low number of cells within the minor sub-
populations of PBMCs make it difficult to
perform cell type–specific eQTL analysis.
Indeed, less than 15 cell types were inves-
tigated in these studies for cell type–spe-
cific eQTL analysis, which is fewer than a
recent study using a bulk RNA-seq–based
approach with 28 immune cell populations
( 11 ). For example, plasmablasts were iden-tified as an immune cell that exhibits sub-
stantial overlap with SLE GWAS top hits
and cell type–specific eQTLs ( 11 ); however,
this signal was not captured by Yazar et
al. and Perez et al., likely because of low
numbers of plasmablasts for analysis.
Moreover, the insufficient resolution of T
cell subclustering limited investigation of
immunologically meaningful T cell sub-
sets, such as regulatory T cells, which are
a small fraction of CD4+ T cells but play
critical roles in regulating autoinflamma-
tory diseases. To include those minor pop-
ulations in the analysis, larger numbers of
cells per donor or enrichment of those cell
types are required for scRNA-seq.
There are several ways to expand the ca-
pacity of single-cell functional genomics.
scRNA-seq data can be used to reconstruct
gene regulatory networks (GRNs) for cell
types or cell lineages by integrating coex-
pression matrices and RNA velocity ( 12 ).
GRNs can also be inferred by integrating
chromatin accessibility data ( 13 ). Recent
advances in single-cell technology make it
possible to jointly profile messenger RNA,
protein, and chromatin accessibility ( 14 ).
This single-cell multi-omics approach pro-
vides further layers of information that can
improve causal GRN inference. A promis-
ing avenue to understand functional fea-
tures of genetic susceptibility is to inter-
rogate specific responses to stimulation
using eQTLs. Given that immune cells can
dynamically change their characteristics
in response to external stimuli and that
disease-associated eQTL effects can be
context specific, genetic effects could be
observed in each context. Thus, interrogat-
ing immune cell responses under differ-
ent stimulation conditions may potentiate
the detection of eQTL effects that may not
be apparent at steady state. The advance-
ment of single-cell technology will fur-
ther expand the application of functional
genetics. Integration of scRNA-seq data
with available functional genetic resources
could pave the way for our understanding
of causal mechanisms of complex diseases. jREFERENCES AND NOTES- S. Yazar et al., Science 376 , eabf3041 (2022).
- R. K. Perez et al., Science 376 , eabf1970 (2022).
- K. K. Farh et al., Nature 518 , 337 (2015).
- E. Choy et al., PLOS Genet. 4 , e1000287 (2008).
- E. E. Schadt et al., Nature 422 , 297 (2003).
- T. Lappalainen et al., Nature 501 , 506 (2013).
- GTEx Consortium, Science 348 , 648 (2015).
- L. R. Lloyd-Jones et al., Am. J. Hum. Genet. 100 , 371
(2017). - W. J. Housley et al., Sci. Transl. Med. 7 , 291ra93 (2015).
- M. G. P. van der Wijst et al., Nat. Genet. 50 , 493 (2018).
- M. Ota et al., Cell 184 , 3006 (2021).
- X. Qiu et al., Cell Syst. 10 , 265 (2020).
- V. K. Kartha et al., bioRxiv 10.1101/2021.07.28.453784
(2021). - E. P. Mimitou et al., Nat. Biotechnol. 39 , 1246 (2021).
10.1126/science.abq0 426
Cell type–specific eQTL analyses Dynamic eQTL analysisGene regulatory network analysis
Cell type–specific network inferenceCell type XCell type–specific eQTL Disease-specific eQTLCell type YscRNA-seqPseudotime (trajectory) RNA velocityGene AGene AGene Y Gene XGene ACis-eQTLAAAAAAAAAA
AAAAAAAAAAAAAAA
AAAAATrans-eQTLStimulation-responsive eQTL8 APRIL 2022 • VOL 376 ISSUE 6589 135Single-cell technology applied to functional genomics
Population-scale single-cell RNA sequencing (scRNA-seq) analyses have the potential to perform multiple
expression quantitative trait locus (eQTL) analyses in a cell type–specific manner. scRNA-seq can be used for
pseudotime-trajectory and RNA-velocity analyses, which can reconstruct cell type–specific gene regulatory
networks. By integrating cell type–specific genetic eQTL effects and eQTLs in response to stimuli, personalized
cell type–specific regulatory networks can be inferred.