Ancient raindrops and networking brain cells

Latest news from Nature 29 March 2012

This press release contains:

---Summaries of newsworthy papers:

Genomics: Cancer therapy gets personal

Comment: Many preclinical cancer findings are not reproducible

Climate science: Coral correlates ice-sheet collapse and sea-level changes

Oncology: New target for cancer therapy

Geoscience: Evidence against the chondritic Earth hypothesis

Developmental biology: DNA methylation in the early embryo

Neuroscience: Brain cells show good networking skills

Physics: Tearing holes in ultrafast light modulation

And finally... Ancient raindrops predict Earth’s early atmosphere

---Mention of papers to be published at the same time with the same embargo

---Geographical listing of authors

Editorial contacts: While the best contacts for stories will always be the authors themselves, in some cases the Nature editor who handled the paper will be available for comment if an author is unobtainable. Editors are contactable via Ruth Francis on +44 20 7843 4562. Feel free to get in touch with Nature's press contacts in London, Washington and Tokyo (as listed at the end of this release) with any general editorial inquiry.

Warning: This document, and the Nature papers to which it refers, may contain information that is price sensitive (as legally defined, for example, in the UK Criminal Justice Act 1993 Part V) with respect to publicly quoted companies. Anyone dealing in securities using information contained in this document or in advanced copies of Nature’s content may be guilty of insider trading under the US Securities Exchange Act of 1934.

PICTURES: While we are happy for images from Nature to be reproduced for the purposes of contemporaneous news reporting, you must also seek permission from the copyright holder (if named) or author of the research paper in question (if not).

HYPE: We take great care not to hype the papers mentioned on our press releases, but are sometimes accused of doing so. If you ever consider that a story has been hyped, please do not hesitate to contact us at [email protected], citing the specific example.


[1] & [2] Genomics: Cancer therapy gets personal (pp 603-607 and 570-575; N&V)

Two studies that may provide a useful resource for efforts to develop personalized cancer therapies are published in Nature this week. By performing large-scale screens of human cancer cell lines, researchers identify potential drug-sensitivity biomarkers to a broad range of cancer drugs.

Levi Garraway and colleagues describe a compilation of genomic data from human cancer cell lines together with pharmacological profiling of 24 drugs across nearly 500 of these lines. The resulting collection, called the Cancer Cell Line Encyclopedia, uncovers both known and new candidate biomarkers that predict response to therapeutic agents. A separate screen of over 600 cancer cell lines with 130 drugs performed by Mathew Garnett and co-workers also identifies genetic signatures associated with drug sensitivity. Of note, they discover a genomic alteration found in Ewing’s sarcoma (a rare childhood cancer) that is associated with sensitivity to a group of pharmacological agents called PARP inhibitors. These observations raise the possibility that PARP inhibitors could be of value in the treatment of Ewing’s sarcoma.

Taken together, these reports offer insights into the genetic basis of sensitivity and resistance to cancer treatment. The identification of potential biomarkers that might predict drug response will hopefully help improve the patient selection process for targeted therapies and drive the development of new therapeutic strategies.


Levi Garraway (The Broad Institute of Harvard and MIT, Cambridge, MA, USA) Author paper [1]
Tel: +1 617 632 6689; E-mail: [email protected]

Mathew Garnett (Wellcome Trust Sanger Institute, Hinxton, UK) Author paper [2]
Tel: +44 1223 494878; E-mail: [email protected]

John Weinstein (University of Texas MD Anderson Cancer Center, Houston, TX, USA) N&V author
Tel: +1 713 563 9296; E-mail: [email protected]


Comment: Many preclinical cancer findings are not reproducible (pp 531-533)

Preclinical cancer research is facing systemic problems ― including the fact that most landmark early findings are not reproducible, according to a validation experiment described by C. Glenn Begley and Lee M. Ellis in a Comment in this week’s Nature.

It has long been known that most compounds that initially seem promising never make it into the clinic. So, over the past decade, before pursuing a particular line of research, scientists at the haematology and oncology department at the Californian biotechnology firm Amgen attempted to confirm published findings related to that work. Out of 53 articles that described something completely new, such as fresh approaches to targeting cancers or alternative clinical uses for existing therapeutics, scientific findings were confirmed in only 6 cases (11%). “Even knowing the limitations of preclinical research, this was a shocking result,” write Begley and Ellis.

Although it remains difficult to mimic the complexity of human cancers in cell lines and animal models, there are ways to increase the robustness of published preclinical cancer research, the authors note. To start, they recommend more thoroughness in the literature, such as blinding all investigators to the control and treatment arms, repeating critical experiments, and including all data in the final publication, including when data on a compound does not completely fit the original hypothesis. They also call for a change in the culture of science that would make it easier for scientists to present data that don’t tell a ‘perfect’ story. This would help researchers to avoid submitting selective datasets for publication, or even massaging data to fit the underlying hypothesis being tested.

“Improving preclinical cancer research to the point at which it is reproducible and translatable to clinical-trial success will be an extraordinarily difficult challenge,” they write. “However, it is important to remember that patients are at the centre of all these efforts.”


C. Glenn Begley (Amgen, Thousand Oaks, CA, USA)
E-mail: [email protected]


[3] Climate science: Coral correlates ice-sheet collapse and sea-level changes (pp 559-564; N&V)

Coral samples collected from Tahiti provide a detailed record of the magnitude and timings of sea-level rise during a controversial episode in the last deglaciation. The findings are reported in Nature this week. Studying sea-level changes provides information that may explain how ice sheets respond to factors that drive climate changes.

A rapid rise in sea level is known to have occurred during a meltwater pulse around 14,500 years ago, but precise details of this event have remained obscure. Pierre Deschamps and colleagues reconstruct sea-level changes by analysing samples of coral collected from reefs in Tahiti; dating shallow-living corals allows the authors to determine the extent of the meltwater pulse and when it happened.

The record establishes that a pulse in melt water occurred between 14,650 and 14,310 years ago, coinciding with a warming event. Sea-level rise was probably between 14 and 18 metres. The authors note that such a large sea-level rise suggests that ice-sheet collapse in Antarctica may have contributed to these changes, previously a point of much contention.


Pierre Deschamps (CEREGE and Université Aix-Marseille, Aix-en-Provence, France)
Tel: +33 6 30 33 31 28; E-mail: [email protected]

Robert Kopp (Rutgers University, Piscataway, NJ, USA) N&V author
Tel: +1 732 200 2705; E-mail: [email protected]


[4] Oncology: New target for cancer therapy (pp 608-612)

A factor that has an important role in a subset of cancers defined by deregulated expression of an oncogene called MYC is identified in this week’s Nature. MYC is difficult to target therapeutically, but an associated protein called ARK5 may represent a potential target for drugs to treat tumours that over express MYC.

MYC encourages cell growth and reproduction and alters cellular metabolism, effects that can contribute to the development of cancer. Martin Eilers and colleagues perform a screen to try to find out what factors help MYC tumours to survive. Their work reveals that MYC-driven cancer cells aredependant on the enzyme ARK5, which helps to alleviate metabolic stress caused by MYC activity. Loss of ARK5 means that the MYC tumours have to endure metabolic stress, which limits their ability to survive.

The authors show that Ark5 is critical in a Myc-driven mouse model of liver cancer. These findings demonstrate that targeting ARK5 or other factors involved in regulating cellular metabolism may provide valid therapeutic strategy to eliminate tumour cells that express deregulated MYC.


Martin Eilers (University of Würzburg, Germany)
Tel: +49 931 318 4111; E-mail: [email protected]


[5] Geoscience: Evidence against the chondritic Earth hypothesis (pp 553-558)

A review of evidence challenging a fundamental assumption about the composition of the Earth is presented in a Perspective in this week’s Nature. The authors argue that the matter from which the Earth formed did not have the same composition as chondritic meteorites, as is commonly assumed. This may have implications for how we interpret the geochemistry of the Earth, its growth, and mantle dynamics.

Until recently, researchers generally thought that the Earth and the other planets of the Solar System should share the same composition as the Sun. However, the chemical make-up of the Earth differs from that of the Sun (as inferred from chondritic meteorites, which are objects from the early Solar System that are presumed to reflect their parent-body composition). In their Perspective, Ian Campbell and Hugh O’Neill discuss hypotheses trying to explain why the composition differs, such as collisions depleting the Earth of chondritic matter late in planet formation. The authors suggest that these theories are missing an important assumption — that the Earth is in fact not chondritic after all.

Campbell and O’Neill conclude that the only way to determine the composition of the Earth reliably is to analyse samples from the Earth’s earliest and most primitive mantle. They hope that the element ratios in these samples may reveal whether they match those of chondritic meteorites or those that would be expected if collisions had caused erosion during the Earth’s formation.

Ian Campbell (Australian National University, Canberra, Australia)
Tel: +61 2 6 125 5159; E-mail: [email protected]


[6] Developmental biology: DNA methylation in the early embryo (AOP)
DOI: 10.1038/nature10960

Genome-scale maps of DNA methylation during early mouse development are presented in a Nature paper. The study provides a high-resolution view of DNA methylation patterns at key developmental stages.

DNA methylation is an epigenetic modification that reflects gene expression patterns and is highly dynamic during early embryonic development in mammals. Alexander Meissner and colleagues profiled DNA methylation across the genome in mouse sperm, eggs and fertilized embryos at several developmental stages. The maps provide a timeline of changes in DNA methylation over critical phases of early embryonic development and a high-resolution view of key genomic targets of DNA methylation. These findings expand on and clarify earlier models of DNA methylation dynamics which were mainly based on immunohistochemistry.


Alexander Meissner (Broad Institute of MIT and Harvard, Cambridge, MA, USA)
Tel: +1 617 495 8801; E-mail: [email protected]


[7] Neuroscience: Brain cells show good networking skills (AOP)
DOI: 10.1038/nature10959

Insights into the generation of non-neuronal cells in the brain, known as glial cells, are reported in Nature this week. The study identifies a sub-population of these cells called astrocytes, which represent a major source of new glial cells during development. Understanding how astrocytes are produced is of great importance as defects in this process are associated with severe neurological disorders, such as brain tumours.

Research in rodents has shown that the population of brain astrocytes increases 6- to 8-fold within weeks of birth, but little is known about the mechanisms underlying this expansion. Astrocytes have a critical role in regulating neuronal cell connections during postnatal development, and have been thought to arise from a distant source of precursor cells. However, Lilly-Yeh Jan and colleagues find evidence that local cell proliferation may contribute to increased astrocyte generation. They show that astrocytes rapidly multiply in the brain during development and these new cells can integrate into existing cellular networks. These findings expand on the previous notion that only injury can prompt local proliferation to occur.


Lily Yeh Jan (University of California, San Francisco, CA, USA)
Tel: +1 415 476 8748; E-mail: [email protected]


[8] Physics: Tearing holes in ultrafast light modulation (pp 580-583; N&V)

An experiment demonstrating the generation of laser light at a series of well-defined frequencies is described in Nature this week. The effect is achieved by irradiating a semiconductor, which leads to the emission of photons at a sequence of different frequencies. These findings may point to a new strategy for vastly increasing the speed of optical communications.

Intense laser fields can rip electrons from atoms and these electrons can then re-collide with the charged atom, resulting in emission of energetic photons. Mark Sherwin and colleagues observe a similar effect between electrons and ‘holes’ in semiconductor wells. They demonstrate that these electrons and holes can be pulled apart and forced to re-collide using an intense terahertz laser field. This results in the production of photon emissions at new frequencies around that of an infrared laser.

The authors conclude that their findings suggest a new mechanism for ultrafast modulation of light, which has potential applications in high-speed optical communications.


Mark Sherwin (University of California, Santa Barbara, CA, USA)
Tel: +1 805 893 3774; E-mail: [email protected]

Benjamin Zaks (University of California, Santa Barbara, CA, USA) co-author
Tel: +1 805 893 4707; E-mail: [email protected]

Rupert Huber (University of Regensburg, Germany) N&V author
E-mail: [email protected]


[9] And finally... Ancient raindrops predict Earth’s early atmosphere (AOP; N&V)
DOI: 10.1038/nature10890

An analysis of fossil imprints of raindrops that fell 2.7 billion years ago in South Africa reveals important clues about Earth’s early atmosphere. The findings, published in Nature this week, suggest that the atmosphere during this time was no more than twice as dense as modern levels, lower than previously hypothesized.

Raindrop imprints serve as a proxy for air density at different periods in Earth’s history because the shape of raindrops is affected by air pressure. Sanjoy Som and colleagues use fossilized raindrop imprints in 2.7-billion-year-old volcanic ash to determine the air density when they fell. At that time, a fainter Sun was nevertheless accompanied by a warm climate. Theories to explain this paradox suggest that Earth’s early atmosphere was substantially thicker to provide a stronger greenhouse effect. However, the present findings suggest that the atmosphere was no more than twice as dense as modern levels, rendering the enhanced greenhouse proposal an unlikely explanation.

As part of their research, the authors performed experiments to define a relationship between imprint size and raindrop properties to allow them to interpret the fossil imprints. They looked at differences in imprints formed by drops on fresh ash (from the 2010 Eyjafjallajökull eruption in Iceland) and weathered ash from around 2.6 billion to 11,700 years ago.


Sanjoy Som (University of Washington, Seattle, WA, USA)
Tel: +1 650 604 1483; E-mail: [email protected]

Paul Renne (University of California, Berkeley, CA, USA) N&V author
Tel: +1 510 644 1350; E-mail: [email protected]



[10] Observation of a roton collective mode in a two-dimensional Fermi liquid (pp 576-579)


[11] The anti-Shine-Dalgarno sequence drives translational pausing and codon choice in bacteria
DOI: 10.1038/nature10965

[12] Crystal structure of a membrane-embedded H+-translocating pyrophosphatase
DOI: 10.1038/nature10963

[13] Differential positioning of adherens junctions is associated with initiation of epithelial folding
DOI: 10.1038/nature10938



The following list of places refers to the whereabouts of authors on the papers numbered in this release. For example, London: 4 - this means that on paper number four, there will be at least one author affiliated to an institute or company in London. The listing may be for an author's main affiliation, or for a place where they are working temporarily. Please see the PDF of the paper for full details.

Canberra: 5

Linz: 10

Hong Kong: 8

Aalto: 10

Aix-en-Provence: 3
Grenoble: 10
Paris: 2

Berlin: 4
Braunschweig: 4
Frankfurt am Main: 13
Hannover: 4
Marburg: 4
Würzburg: 4

Milan: 4

Kashiwashi: 3
Saitama: 7
Tokyo: 3
Yokosuka: 3

Amsterdam: 4

Taipei: 12
Taiwan: 12

Lausanne: 2

Cambridge: 2
Edinburgh: 9
Hinxton: 2
Norwich: 2
Oxford: 3

Emeryville: 1
La Jolla: 7
Moffett Field: 9
San Diego: 1
San Francisco: 7, 11
Santa Barbara: 8
Chicago: 13
Chevy Chase: 1, 2
Boston: 1, 2
Cambridge: 1, 2, 6
Charlestown: 2
New Jersey
Princeton: 13
New York
Buffalo: 10
New York: 1
Oak Ridge: 10
Seattle: 1, 9



From North America and Canada

Neda Afsarmanesh, Nature New York
Tel: +1 212 726 9231; E-mail: [email protected]

From Japan, Korea, China, Singapore and Taiwan

Eiji Matsuda, Nature Tokyo
Tel: +81 3 3267 8751; E-mail: [email protected]

From the UK

Rebecca Walton, Nature London
Tel: +44 20 7843 4502; E-mail: [email protected]


About Nature Publishing Group (NPG):

Nature Publishing Group (NPG) is a publisher of high impact scientific and medical information in print and online. NPG publishes journals, online databases and services across the life, physical, chemical and applied sciences and clinical medicine.

Focusing on the needs of scientists, Nature (founded in 1869) is the leading weekly, international scientific journal. In addition, for this audience, NPG publishes a range of Nature research journals and Nature Reviews journals, plus a range of prestigious academic journals including society-owned publications. Online, provides over 5 million visitors per month with access to NPG publications and online databases and services, including Nature News and NatureJobs plus access to Nature Network and Nature Education’s

Scientific American is at the heart of NPG’s newly-formed consumer media division, meeting the needs of the general public. Founded in 1845, Scientific American is the oldest continuously published magazine in the US and the leading authoritative publication for science in the general media. Together with and 15 local language editions around the world it reaches over 3 million consumers and scientists. Other titles include Scientific American Mind and Spektrum der Wissenschaft in Germany.

Throughout all its businesses NPG is dedicated to serving the scientific and medical communities and the wider scientifically interested general public. Part of Macmillan Publishers Limited, NPG is a global company with principal offices in London, New York and Tokyo, and offices in cities worldwide including Boston, Buenos Aires, Delhi, Hong Kong, Madrid, Barcelona, Munich, Heidelberg, Basingstoke, Melbourne, Paris, San Francisco, Seoul and Washington DC. For more information, please go to

Published: 29 Mar 2012

Contact details:

The Macmillan Building, 4 Crinan Street
N1 9XW
United Kingdom

+44 20 7833 4000
News topics: 
Content type: