emerge from the crystalline
phase. The proposed methodol-
ogy is general and applicable
to other physical systems
undergoing first-order phase
transitions. —YS
Proc. Natl. Acad. Sci. U.S.A. 118 ,
2106230118 (2021).
CELL METABOLISM
Probing the TCA cycle
The tricarboxylic acid (TCA)
cycle is key to cellular metabo-
lism, being responsible for
generating ATP, the energy
source for cellular processes.
In addition, the TCA cycle
produces precursors for
synthesizing other biological
molecules, including nones-
sential amino acids. Despite
its central role, the conse-
quences of TCA dysfunction are
poorly understood. Ryan et al.
investigated the consequences
of inhibiting the TCA cycle in
murine kidney epithelial cells by
either inhibiting or ablating two
key enzymes. The authors moni-
tored the effects of disruption
using metabolomics, transcrip-
tomics, and proteomics. An
impaired TCA cycle alters amino
acid and redox metabolism and
activates the integrated stress
response of the cell to rewire
transcription and translation
and compensate for the amino
acid and redox stress. —VV
eLife 11 , e72593 (2021).STEM WORKFORCE
Retention over
recruitment
Retention in academia
remains a persistent barrier to
proportional representation.
Shaw et al. used 25 years of
National Science Foundation
data on the racial and
ethnic composition ofSTEM academics, from
undergraduates to tenured
professors, to construct a
null model of ethnic and
racial representation in the
US science community.
Comparisons between the
null model and current
representation allowed them
to measure the effects of
retention while controlling
for recruitment at different
academic stages. Their results
show that failed retention,
occurring at different
stages depending on race
and ethnicity, negatively
affected Black, Indigenous,
and Hispanic scholars the
most. A substantial drop in
racial/ethnic representation
between students (graduate
and undergraduate) and
researchers (postdocs and
faculty) was found, suggesting
that recruiting diverse students
is not enough, and an increasedfocus on inclusion/retention is
needed. —MMc
PLoS ONE 16 , e0259710 (2021).ORGANIC CHEMISTRY
Copper lights up
acid coupling
Carboxylic acids are relatively
abundant feedstocks for the
preparation of more elaborate
compounds in fine chemical,
pharmaceutical, and agro-
chemical synthesis. Recently,
transformation of these acids
into redox-active esters has
enabled efficient coupling
chemistry using light-activated
catalysts. Li et al. report that
a simple copper salt can take
the place of both the ester
group and the catalyst, single-
handedly oxidizing the acid for
decarboxylative coupling under
blue light. A variety of sulfon-
amides, amides, and alcohols
served as effective coupling
partners. —JSY
Nat. Chem. 10.1038/s41557-021-
00834-8 (2022).CELL TRACKING
Let ELEPHANT do
the work
Tracking individual cells and their
progeny through development is
a labor-intensive and technically
challenging task, and it often
takes months to analyze a single
dataset. Sugawara et al. describe
a setup they call efficient learning
using sparse human annota-
tions for nuclear tracking
(ELEPHANT), which performs
cell tracking in three dimen-
sions (3D) with minimal user
input. ELEPHANT optimizes
incremental deep learning using
sparse annotations to detect
nuclei in 3D and then links these
nuclei through time in 4D image
datasets. ELEPHANT success-
fully tracked cell lineages during
embryonic development in nem-
atodes, during limb regeneration
in a crustacean, as well as in
human intestinal organoids and
breast carcinoma cells. All in all,
data analysis time was reduced
from months to weeks. —SMH
eLife 11 , e69380 (2022). ADAPTED FROM ISTOCK.COM/ERHUI1979 BY M. ATAROD/SCIENCERESEARCH | IN OTHER JOURNALS
280 21 JANUARY 2022 • VOL 375 ISSUE 6578 science.org SCIENCE
SOCIAL MEDIADoes Twitter bend to the Right?
C
oncerns are increasing about political polarization and its potential for causing widespread
social disruption in a world also facing environmental disruption. Political discussions form
the bulk of exchanges on the social media platform Twitter. In 2016, Twitter introduced
machine learning for ranking content tuned to individual preferences. Huszár et al.
investigated whether this new mode of content provision selectively amplifies political
flavors. This long-term experiment includes a control group free of algorithmic personalization. In
tweets made by elected politicians from six of the seven countries studied, it appears the political
Right had a louder voice than any other persuasion. Meanwhile, Levin et al. introduce recent
interdisciplinary research, drawing inspiration from evolutionary theory and systems science,
to explore the influences reshaping global politics in potentially dangerous directions. —CA
Proc. Natl. Acad. Sci. U.S.A. 119 , e2025334119 (2022); 118 , e2116950118 (2021).