INSIGHTS | PERSPECTIVES
science.org SCIENCEtor of TaCOL-B5. According to functional
analysis of the conserved domains, the au-
thors postulate that the TaCOL-B5 transcrip-
tion factor modulates multiple traits, such as
flowering time and plant height, through dif-
ferent conserved domains.
The next challenge is to untangle these dif-
ferent yield-related traits through targeted
modification of specific domains or specific
amino acids. Further studies into TaCOL-B5,
its conserved domains, its role in growth reg-
ulation networks, and its responses to diverse
environmental cues can help to fine-tune
wheat cultivars to the specific needs of grow-
ers worldwide. Such research should also
include other regulators of flowering time
and plant architecture, such as FLOWERING
LOCUS-T family members ( 7 ). Fundamental
knowledge of the underlying molecular-
genetic networks provides opportunities
for generating new variation and increasing
yield potential through knowledge-driven
breeding as well as by genetic modifications
and gene editing ( 2 , 3 , 7 , 11 ).
Rapid climate change, reduced resources,
and biotic and abiotic stress ( 5 , 6 ) call for a
multidisciplinary approach to tackle these
challenges. Innovations and technologies,
such as genomic selection, gene editing,
precision agriculture, and intercropping
(growing at least two different crops in a
field at the same time), as well as advanced
phenotyping technologies are needed to ap-
ply scientific knowledge of plant develop-
ment and plasticity to breeding and grow-
ing practices ( 7 , 12 – 14 ). Nonetheless, the
introduction of genes or alleles into new
varieties that increase the yield potential
of cereals is a major goal for plant breed-
ers and scientists to enable sustainable crop
production. The identification of TaCol-B5
by Zhang et al. offers a new route to maxi-
mize yield in wheat. j
REFERENCES AND NOTES
- M. Haas, M. Schreiber, M. Mascher, J. Integr. Plant Biol.
61 , 204 (2019). - D. F. Calderini et al., New Phytol. 230 , 629 (2021).
- X. Zhang et al., Science 376 , 180 (2022).
- J. Putterill, F. Robson, K. Lee, R. Simon, G. Coupland, Cell
80 , 847 (1995). - Food and Agriculture Organization of the United
Nations (FAO), “The impact of disasters and crises on
agriculture and food security: 2021” (2021); https://doi.
org/10.4060/cb3673en. - N. Alexandratos, J. Bruinsma, “World Agriculture
Towards 2030/2050: The 2012 Revision” (FAO, ESA
Working Paper no. 12-03, 2012); http://www.fao.org/3/
ap106e/ap106e.pdf. - Y. Eshed, Z. B. Lippman, Science 366 , eaax0025 (2019).
- International Wheat Genome Sequencing Consortium
et al, Science 361 , eaar7191 (2018). - P. Borrill, New Phytol. 228 , 1721 (2020).
- A. Watson et al., Nat. Plants 4 , 23 (2018).
- L. Liu et al., Nat. Plants 7 , 287 (2021).
- C. Li et al., Nat. Plants 6 , 653 (2020).
- D. Sun, K. Robbins, N. Morales, Q. Shu, H. Cen, Tr e n d s
Plant Sci. 27 , 191 (2022). - C. Uauy, Curr. Opin. Plant Biol. 36 , 142 (2017).
10.1126/science.abo7429
GENOMICSPopulation genetics meets
single-cell sequencing
S ingle-cell technology can be used to understand
the genetic basis of human diseases
By Tomokazu S. Sumida1,2 and
David A. Hafler1,2,3U
sing high-throughput genotyping
platforms to identify single-nucle-
otide polymorphisms (SNPs) has
allowed genome-wide association
studies (GWASs) to generate an unbi-
ased classification of human diseases
that are associated with common genetic
variation. There are many common allelic
variants in noncoding regions that have
small effect sizes with complex interactions
that are highly cell type and cell state de-
pendent. Moreover, progress toward un-
derstanding disease mechanisms has been
limited by the challenges of assigning mo-
lecular function to most GWAS “hits” that
are noncoding sequences associated with
disease. On pages 154 and 153 of this issue,
Yazar et al. ( 1 ) and Perez et al. ( 2 ), respec-
tively, use multiplexed single-cell RNA se-
quencing (scRNA-seq) with fine mapping
of autoimmune disease–associated genetic
variants to provide a resource that allows
the large-scale identification of genotype-
phenotype interactions. Notably, these two
studies provide a comprehensive catalog of
immune cell profiles that opens the door to
a new era of functional genetics.
Most GWAS variants associated with dis-
eases map to noncoding regions that are
highly enriched in regulatory elements,
indicating that those variants are likely
to exert their effects through the modu-
lation of gene expression ( 3 ). Expression
quantitative trait locus (eQTL) analysis is
used to measure the association between
genetic variants and gene expression. This
requires RNA expression to be averaged
across bulk populations of cells ( 4 ), allow-
ing the characterization of genetic vari-
ants that are significantly associated with
gene expression in a population sample
( 5 – 8 ). Previous integration of bulk RNA-
seq–based eQTL analysis with tissue- orcell type–specific gene expression profiles
provided further evidence that eQTL ef-
fects act in a tissue- and cell type–specific
manner, although this approach is biased
to known cell types and does not allow the
identification of cell subsets and transi-
tional states. Although scRNA-seq coupled
with GWAS to identify disease-associated
variants has the potential of revealing both
individual cell types and states where vari-
ants exert their effects, the cost and low
throughput of scRNA-seq had not allowed
this approach at the necessary scale.
By overlaying likely causative SNPs onto
maps of histone modifications (which reg-
ulate gene expression), such as the histone
3 Lys^27 (H3K27) acetylation maps of differ-
ent cell types from the ENCODE project,
cells that are likely influenced by disease-
associated gene variants can be identi-
fied ( 3 ). Although the effect size of allelic
variants can be small, functional analysis
of single haplotypes (a group of alleles of
different genes that are inherited together)
revealed that the biologic effects could
be substantial. For example, risk variants
associated with the autoimmune disease
multiple sclerosis that were proximal to
the nuclear factor κB subunit 1 (NFKB1)
gene were associated with increased patho-
genic NFκB signaling with tumor necrosis
factor–α stimulation in healthy individu-
als ( 9 ). Although studies that investigated
specific genotype-phenotype interactions
were informative, they did not provide the
necessary scalability to broadly examine
hundreds of allelic variants.
To overcome those limitations, high-
throughput scRNA-seq was previously used
to conduct a cell type–specific eQTL study
at subpopulation scale as a proof of con-
cept ( 10 ). Cell type–specific eQTL analysis
was performed with six immune cell types
identified from ~25,000 peripheral blood
mononuclear cells (PBMCs) extracted from
45 individuals. The study confirmed that
scRNA-seq–based eQTL analysis can rep-
licate observations of previously known
“local” eQTLs that act in cis to modulate
gene expression (cis-eQTLs). This study
showed the utility of scRNA-seq for inte-
grating multiple sets of immune cell tran-(^1) Department of Neurology, Yale School of Medicine, New
Haven, CT, USA.^2 Broad Institute of Massachusetts Institute
of Technology and Harvard University, Cambridge, MA,
USA.^3 Department of Immunobiology, Yale School of
Medicine, New Haven, CT, USA. Email: tomokazu.sumida@
yale.edu; [email protected]
134 8 APRIL 2022 • VOL 376 ISSUE 6589