INSIGHTS | PERSPECTIVES
1122 17 MARCH 2017 • VOL 355 ISSUE 6330 sciencemag.org SCIENCE
the developmental processes that parallel
human language acquisition. It may be no
accident that the main evidence for vocal
plasticity and the role of parents in shaping
vocal structure and turn-taking has been
seen in primates that breed cooperatively
(marmosets and tamarins). In these species,
most group members assist with infant care,
a system thought to parallel human child-care. However, no studies have yet described
vowel-like sounds in these monkeys, so marmosets and tamarins may be useful primarily for developmental studies.
It is probable that early humans faced
evolutionary pressures that differed from
those encountered by other primates and
that have made our complex communication system adaptive. Language may have
been important for coordinating activities
in large cooperative groups in which different individuals played different roles, as
was likely among our early ancestors. Neurological and genetic specializations likely
allowed humans to develop articulatory
and sequencing abilities that are not seen in
monkeys and apes. There may also be cognitive and motivational limits: If individuals
can thrive without complex vocal signaling,
there would be little motivation to push the
communication further. Finally, different
sensory and motor systems may be important. We tend to evaluate language through
a vocal/auditory system, whereas research
on apes is beginning to illustrate the complexity of gestural communication (12).
Nonhuman primates do not talk, but we
should not expect them to. Each species has
its own adaptations for communication.
Nevertheless, there is much about language
evolution that we can learn from nonhuman primates, provided that we study a
variety of species and consider the multiple
components of speech and language. j
1. V.M.Janik, P.J.B.Slater, Adv. Stud Behav.26,59(1997).
2. P. H. Lieberman, D. H. Klatt, W. H. Wilson, Science 164, 1185
3. A. P. Melis, P. Grocke, J. Kalbitz, M. Tomasello, Psych.Sci .
27, 987 (2016).
4. C. T.Snowdon, Adv. Stud. Behav.40,239(2009).
5. D. Y. Takahashi et al., Science 349,734(2015).
6. M. H. Goldstein, A. P. King, M. J. West, Proc. Natl. Acad. Sci,
U.S.A. 100, 8030 (2003).
7. W. T.Fitch, V.B.deBoer, N.Mathur, A.A.Ghazanfar, Sci.Adv.
2, e1600723 (2016).
8. L.-J. Boe et al., PLOS ONE 12, e0169321 (2017).
9. S. E. Calcutt, E. V. Lonsdorf, K. E. Bonnie, M. S. Milstein, S. R.
Ross, Behaviour 151, 1967 (2014).
10. K.A.Cronin, Anim. Behav.84,1085(2012).
11. S. C. Levinson, Trends Cogn. Sci .20, 6 (2016).
12. M. Fröhlich et al ., Sci. Rep. 6, 25887 (2016).
13. D. Y. Takahashi,A.R.Fenley, A.A.Ghazanfar, Philos. Trans.
R. Soc. B. 371, 20150370 (2016).
14. C. P. Chow, J. F. Mitchell, C. T. Miller, Proc. R. Soc. B .282,
15. A. R. Pfennig et al ., Science 346, 125684 (2014).
By Armin Scheben and David Edwards
The global population is expected to rise from 7.3 billion to 9.7 billion by 2050 (1). At the same time, climate change poses increasing risks to crop roduction through droughts and pests (2). Improved crops are thus urgently needed to meet growing demand for
food and address changing climatic conditions. Genome-editing technologies such as
the CRISPR (clustered regularly interspaced
short palindromic repeat)/Cas (
CRISPR-associated protein) system (3) show promise
for helping to address these challenges, if the
precision of genome editing is improved and
the technology is approved and accepted by
regulators, producers, and consumers.
From 1981 to 2000, rice, maize, and wheat
varieties that had been improved through
traditional plant breeding boosted crop yields
by 22 to 46% in Asia and Latin America (4).
To meet growing demand by 2050, however,
a global increase in crop production of 100
to 110% from 2005 levels is required (5). At
the same time, climate change is predicted to
lower regional crop yields, especially in wheat
and maize. In semi-arid developing countries
such as Brazil and Argentina, major crop
yields may decline by up to 30% by 2030, and
in sub-Saharan Africa, yields may decline by
22%, with losses of more than 30% in South
Africa and Zimbabwe (6).
Traditional plant breeding is based on
crossing germplasm and then selecting
individuals with desirable traits (see the
figure). Although this approach has been
extraordinarily successful, it can take more
than 10 years, and in some cases decades,
to develop an improved variety. Genomic
tools can improve selection efficiency, but
breeding remains laborious and dependent
on shuffling existing diversity. Given the food
security concerns that the human population
faces, scientists are turning to genome
editing approaches such as CRISPR/Cas (see
the figure). Advantages of genome editing
over conventional and earlier transgenic
approaches are the low cost, ease of use,
lack of transgenes permanently introduced
into crop germplasm, and the high level of
multiplexing (editing of multiple targets)
possible (7). The latter allows rapid trait
stacking and editing of gene networks in
their native context to improve quantitative
traits such as drought tolerance and yield.
Multiplexing is particularly useful in
polyploid crops such as wheat (which have
more than two sets of chromosomes) because
it allows simultaneous editing of multiple
gene copies. Furthermore, many simple trait
improvements involving few genes have
likely already been made in staple crops,
so that trait stacking and more complex
modification of gene networks is required to
further enhance global yields. The low costs
and ease of use of genome editing may also
facilitate improvement of subsistence crops
such as cassava, with potentially substantial
yield increases in sub-Saharan Africa and
The CRISPR/Cas system consists of a
guide RNA containing a target sequence of
usually 20 nucleotides and a Cas nuclease
such as the commonly used Cas9, which
cleaves double-stranded DNA at the target
site. CRISPR/Cas can induce mutations at
virtually any genomic site in any organism,
functioning like a find-and-replace tool in a
word processor. Insertion and/or deletion of
nucleotides at the target site occur because of
DNA repair errors, whereas specific insertions
are achieved by providing template DNA. For
plant breeding, this means that scientists
can edit the genomes of elite varieties to
produce new varieties in a single generation,
unconstrained by existing variation and
without having to select for favorable
combinations of alleles in large populations.
However, unlike traditional breeding, such
targeted genome editing requires knowledge
of the nucleotide sequence and function
of the target to design the guide RNA and
predict the editing outcome.
Genome editors take on crops
Genome editing technologies may help to enhance
global food security
School of Biological Sciences and Institute
of Agriculture, University of Western Australia, Perth, WA,
Australia. Email: firstname.lastname@example.org;
“Improved crops are ...
urgently needed to meet
growing demand for food
and address changing