SCIENCE science.orgevidence against the chiral spin-triplet pic-
ture was collected ( 13 – 15 ). This paradigm
change also implied that the Zeeman split-
ting could indeed break pairs, and thus the
stage could be ideal for an FFLO state.
Kinjo et al. found signatures of an FFLO
state, a result in line with the dismissal of
the chiral spin-triplet picture. By measur-
ing the nuclear magnetic resonance (NMR)
signal at specific sites of the crystal lat-
tice, they measured the spatial variation
in spin density at 70 mK—well below the
superconducting critical temperature. The
authors gradually increased the magnetic
field while making sure that the entire vol-
ume of the sample remained in the super-
conducting state up until the critical field
of 1.4 T. At low magnetic fields, Sr 2 RuO 4 be-
haved like a conventional superconductor,
with a single NMR peak accompanied by a
resonance frequency shift, which is known
as Knight shift. This shift was smaller in the
superconducting phase than in the normal
phase, which is in line with the expecta-
tions for conventional superconductivity.
However, approaching the critical magnetic
field, a second peak appeared in the NMR
signal. The associated Knight shift was
larger than in the normal phase, which is
something that cannot be easily explained
by a mixture of normal and superconduct-
ing phases and instead could be consistent
with spatial spin-density modulations asso-
ciated with an FFLO state.
The discovery of a possible FFLO state
in the layered perovskite Sr 2 RuO 4 paves the
way for unprecedented studies of the elu-
sive superconducting state. The evidence
presented by Kinjo et al. is compelling but
still indirect. The smoking gun—the direct
measurement of the spatial modulations of
the superconducting order parameter—re-
mains at large. For now, Sr 2 RuO 4 confirms
itself as a system of neverending wonders.
It might not be a chiral spin-triplet super-
conductor, but it is certainly one of a kind.
More surprises could be in store for the
coming years. j
REFERENCES AND NOTES
- J. Bardeen et al., Phys. Rev. 108 , 1175 (1957).
- P. Fulde, R. A. Ferrell, Phys. Rev. B 135 , A550 (1964).
- A. I. Larkin et al., Zh. Exp. Teor. Fiz. 47 , 1136 (1964).
- K. Kinjo et al., Science 376 , 397 (2022).
- H. A. Radovan et al., Nature 425 , 51 (2003).
- A. Bianchi et al., Phys. Rev. Lett. 91 , 187004 (2003).
- K. Cho et al., Phys. Rev. B 79 , 220507R (2009).
- H. Mayaffre et al., Nat. Phys. 10 , 928 (2014).
- G. Zhang et al., Phys. Rev. Lett. 116 , 106402 (2016)
- T. M. Rice, M. Sigrist, J. Phys. Condens. Matter 7 , L643
(1995). - K. Ishida et al., Nature 396 , 658 (1998).
- S. Yonezawa, T. Kajikawa, Y. Maeno, Phys. Rev. Lett. 110 ,
07703 (2013). - A. Pustogov et al., Nature 574 , 72 (2019).
- K. Ishida et al., J. Phys. Soc. Jpn. 89 , 034712 (2020).
- A. N. Petsch et al., Phys. Rev. Lett. 125 , 217004 (2020).
10.1126/science.abn3794CANCERA fresh look at somatic
mutations in cancer
A nalysis of cancer genome sequences reveals
new mutational signatures
By Dávid SzütsS
equencing the genomic DNA of tu-
mor samples has provided rich data
on mutagenic processes and revealed
that information derived from heter-
ogeneous mutational landscapes of
cancers can complement the identi-
fication of driver gene mutations in aiding
reliable personalized diagnosis and treat-
ment selection. On page 368 of this issue,
Degasperi et al. ( 1 ) present a comprehen-
sive analysis of the patterns of somatic base
substitution mutations in human cancers.
Their work builds on landmark studies
that revealed that mutational catalogs, or
spectra, can yield constituent mutational
components that reflect biological pro-
cesses ( 2 – 4 ). Combining component “sig-
natures” can adequately reconstitute the
mutation spectrum of individual cancer
samples and facilitate an understanding of
the underlying mutagenic mechanisms.
Mutational signatures categorize muta-
tions according to the sequence context
and the type of change, which are generally
specific to an external mutagen or endog-
enous mutagenic process. There are two
fundamental approaches in mutational
signature analysis: using unsupervised
algorithms to define de novo signatures
in a dataset and deconstructing mutation
spectra into predefined signatures. Human
cancers are expected to display a finite va-
riety of mutational processes, which fuels
the endeavor to establish a comprehen-
sive set of cancer-associated mutational
signatures, and their potential clinical use
would benefit from a scientific consensus.
Research in this area has progressed rap-
idly, and by 2021, the Catalogue of Somatic
Mutations in Cancer (COSMIC) listed ~60
real single-base substitution (SBS) signa-
tures and several artifactual ones ( 5 ). The
increasing number of signatures, which
were not all derived from a single analysis,
have made it progressively more complex
to fit signatures to new datasets, with a
judicious preselection of mutational signa-tures often necessary for the most mean-
ingful results.
The study by Degasperi et al. used an
elegant and clear solution for defining mu-
tational signatures, initiated using organ-
specific sample cohorts. A stable two-step
process first found common signatures in
spectra from samples that cluster with oth-
ers by similarity, and then the spectra from
excluded samples were deconstructed us-
ing defined common signatures and poten-
tial additional rare signatures. The organ-
specific signatures were then clustered,
filtered, and averaged to establish 120
reference signatures that present a picture
of general mutational processes in cancer
(see the figure). The palette of common
signatures appears to be saturated, but the
approach leaves the door open to finding
rare signatures.
The greatest asset of the study is the
notable 12,222-sample dataset from the
100,000 Genomes Project of Genomics
England, which has a greater number of
whole-genome sequences than previous
major cancer sequencing projects put to-
gether ( 6 ). The results were validated on
datasets from the International Cancer
Genome Consortium and the Hartwig
Foundation. The Genomics England data-
set contains hundreds of samples per or-
gan type, which allowed the detection of
rare signatures that were present in less
than 1% of each organ set.
What has the new analysis revealed?
Beyond confirming most, but not all, of
the previously identified base substitution
signatures, it added 40 new SBS and 18
double-base substitution signatures to the
reference set. The signatures bear evidence
of the expected mutagenic processes: effects
of cellular metabolism, deficiencies of DNA
repair, and environmental sources. Of par-
ticular interest are signatures that only show
subtle differences but can be tied to different
mechanisms through association with driver
gene mutations, such as inactivating methyl-
CpG binding domain protein 4 (MBD4) mu-
tations that explain signature SBS96, which
is similar to other signatures caused by CpG
deamination ( 7 ). Because the mechanism be-
hind many signatures is still unknown, theInstitute of Enzymology, Research Centre for Natural
Sciences, Budapest, Hungary. Email: [email protected]22 APRIL 2022 • VOL 376 ISSUE 6591 351