Storytelling science

“Peter, we changed its DNA!” —Mira Sorvino as entomologist Dr. Susan Tyler in Mimic

Most practicing scientists cringe when they watch Hollywood versions of science; the above line from Mimic is a memorable example that I find particularly cringe-worthy. Risks of sweeping generalizations aside, I think most bench scientists would agree that science is just not accurately represented in popular culture. How many unrealistic scenes of the molecular forensic lab in CSI does a practicing scientist need to watch to draw this conclusion anyway?

At “Celluloid Science: Humanizing Life in the Lab," an event held at the New York Academy of Science last night, regular science contributor to the New York Times Carl Zimmer led a thoughtful panel discussion on the challenges and efforts of those who tell stories of science through film.

Valerie Weiss, a filmmaker and former biophysicist, kicked off the panel with commentary and clips from her upcoming feature film, Losing Control. The independent comedy produced by her company, PhD Productions, is set for release on Valentine’s Day 2012 and is centered on a female scientist seeking experimental proof that her boyfriend is “the one.” The film is loosely based on Weiss’s experience in graduate school at Harvard University. Having been a female bench scientist for most of my adult life, the trailer seems largely accurate and endlessly amusing. For better or for worse, I can’t say I have (or know anyone who has) personal experience with the ubiquitous laboratory safety shower, but I look forward to watching the scene in its entirety, as I can’t imagine any scene involving unfortunate snafus with laboratory safety showers as not funny.

David Heeger, Professor of Psychology and Neuroscience at NYU, gave a short talk on his research on neurocinematics, which is essentially the science of human reaction to film. Heeger’s experimental approach uses functional MRI (fMRI) to monitor human brain activity while watching film. Heeger presented intriguing results that show that for well-produced films, the resulting brain activity patterns of research subjects are very similar. In contrast, when subjects are shown segments of reality (in this case, clips from a camera set up in Washington Square Park, which some might argue are actually the very opposite of reality), there is a very weak correlation between individuals’ brain activities. The research seems to suggest that film quality might not be subjective, although brain activity does not correlate with film preferences between individuals.

The panel then shifted to a well-known evolutionary biologist from the University of Wisconsin at Madison, Sean Carroll, who is also the Vice President for Education at the Howard Hughes Medical Institute. Carroll spoke on his belief in the power of the story, and HHMI’s mission to reach a broad audience with well-crafted and accurate scientific stories through film. To this end, HHMI has set aside $60 million over 5 years for its in-house film production unit. Carroll screened a short 10 minute film entitled, “The Making of the Fittest: Natural Selection and Adaptation” that is intended for the classroom. The film tells the story of the evolution of light and dark rock pocket mice in the lava flows of the Southwest desert. With wide open landscapes as the backdrop, the film accurately illustrates a specific example of Darwin’s process of natural selection. Carroll summarized the message of the film elegantly, “Mutation is random, but natural selection is not.”

Melanoma sequencing identifies new druggable targets

Next-generation sequencing technologies are enabling unbiased searches for new cancer genes at an unprecedented scale. In 2011, a flurry of cancer exome and whole-genome papers have been published in high-impact journals, with more in the pipeline. The first genes to be targeted for personalized treatment will be ones harboring recurrent mutations at a high frequency and those with already known inhibitors/modulators. The delivery of personalized therapies in cancer will no longer be bottlenecked by a lack of targets; the development of effective therapies will require new insights into how cancers become resistant to drugs and hopefully, therapeutic interventions that bypass acquired resistance.

Melanomas are tumors of melanocytes, the pigment-producing cells in skin, and occasionally in the iris of the eye (uveal melanoma). Every year, melanoma is diagnosed in approximately 160,000 new cases and leads to death in 48,000 people worldwide. Exposure to UV radiation is a major risk factor for melanoma, and the incidence of melanoma is rising. Patients with stage IV disease have a poor prognosis, with a median survival of 8-18 months post-diagnosis. Currently, dacarbazine is the only chemotherapeutic drug approved by the FDA for metastatic melanoma and has a median overall survival of 5.6-7.8 months after treatment begins.

Phase III clinical trial results for PLX4032 show remarkable clinical activity in melanomas with BRAF mutations.

In 2002, mutations in the protein kinase BRAF were identified in 66% of malignant melanoma tumors analyzed; subsequent studies estimate that 40-60% of melanomas have mutations in BRAF. Nearly 90% of mutations in BRAF lead to a valine to glutamate change (p.Val600Glu). In order to develop small molecule inhibitors of p.Val600Glu, Gideon Bollag of Plexxikon, Inc. in Berkeley, CA and colleagues screened a library of 20,000 compounds using a structure-guided approach. In 2008, the scientists reported that the compound PLX4720 specifically inhibits BRAF p.Val600Glu with potent anti-melanoma effects in in vitro and in vivo models.

Recently, the long-awaited results of phase III clinical trials of the BRAF kinase inhibitor (PLX4032, an analogue of PLX4720) were published ), showing increased overall survival in patients on PLX4032 compared to the chemotherapeutic agent dacarbazine. In the study, 675 patients with metastatic melanoma carrying the p.Val600Glu mutation were randomly assigned to treatment with PLX4032 or dacarbazine. After six months, overall survival in the PLX4032 group was 84%, but only 64% in the group receiving dacarbazine. Of all patients treated with PLX4032, 48% had a confirmed response, although the authors report that most patients showed tumor shrinkage. Despite the obvious benefit of PLX4032 treatment, the effectiveness of the treatment is short-lived, as many responsive tumors become resistant to treatment within 8-12 months. Determining the mechanisms responsible for acquired resistance to PLX4032 is a top priority for further investigation, and some insights have already been made. Surprisingly, in three recent studies

resistance has not been found to be caused by second-site mutations in BRAF, but activation of alternative compensatory pathways.

The promise of personalized treatment in cancer is another step closer with the identification of mutations in ERBB4 in ~20% of melanoma tumors by Yardena Samuels and colleagues.

In a study published in this journal, Yardena Samuels of the National Human Genome Research Institute and colleagues sequenced 86 members of the protein tyrosine kinase gene family in 29 metastatic melanomas. They identified 19 genes with 30 somatic mutations and sequenced these 19 genes in a larger number of melanomas. The most highly mutated gene in this screen was ERBB4, with mutations in 19% of melanomas. Crystal structure analysis showed that positioning of mutations found in ERBB4 were reminiscent of mutations found in the ERBB family members EGFR and ERBB2 in lung cancer, glioblastoma and gastric cancer.

Lapatinib is a dual receptor tyrosine kinase inhibitor that targets EGFR and HER2; this drug was approved by the FDA in 2010 for treatment of HER2-overexpressing metastatic breast cancer. Functional analysis of a subset of the ERBB4 mutations led Samuels and colleagues to hypothesize that mutant ERBB4 might be inhibited by the pan-ERBB inhibitor lapatinib. The scientists found that this small molecule can reduce melanoma cell proliferation in vitro. Armed with this pre-clinical validation data, a multi-center phase II clinical trial at the National Institutes of Health Clinic Center (Principal Investigator Dr. Udo Rudloff) and Memorial Sloan-Kettering Cancer Center (Principal Investigator Dr. Mark Dickson) is currently underway. The trial is testing lapatinib for stage IV melanomas harboring ERBB4 mutations (Clinical Identifier NCT01264081 and Protocol 10-222, respectively).

In the first exome analysis of melanoma also published in this journal and available free for download, Samuels and colleagues identified 68 genes with somatic mutations. They found that GRIN2A is mutated in 25% (34 mutations in 135 samples) of melanomas analyzed. GRIN2A is a glutamate (NMDA) receptor subunit that also is mutated in neurodevelopmental phenotypes and reported in this journal last year. The germline mutations in neurodevelopmental disorders do not overlap with the somatic mutations in melanoma, but mutations in both diseases lead to early stop codons that presumably are null alleles. In melanoma, the spectrum of mutations in GRIN2A cluster to two regions of the protein. The authors also observed three recurrent alterations in residues that are evolutionarily conserved.

The glutamate signaling pathway is likely to be functionally significant in melanoma, since another highly mutated gene was PLCB4, a phosopholipase isozyme known to signal through metabotropic glutamate receptors. Adding another connection between phospholipase activity and melanocytic-derived tumors, Boris Bastian and colleagues recently reported recurrent somatic mutations in the G-proteins GNA11 and GNAQ in uveal melanoma. Somatic mutations in GNAQ were also found in blue naevi; mutations in GNAQ were limited to codon 209 and lead to overactivation of the MAP kinase pathway. Both GNA11 and GNAQ are G-proteins (alpha subunits) that stimulate phospholipase C-beta.

In another link between melanoma and glutamate signaling, Samuels and colleagues note that GRM3, a metabotropic glutamate receptor activated by glutamate, is mutated in 16% of melanoma tumors (unpublished data). A role for glutamate signaling has been seen in neuronal tumors, with an excess of glutamate associated with more aggressive growth of tumor cells. There are published reports of glutamate inhibitors suppressing tumor growth; further investigation of the possible therapeutic benefit of modulating glutamate signaling in GRIN2A, PLCB4 and GRM3 melanomas seems like a natural next step.

Finally, Samuels and colleagues identified a recurrent mutation in TRRAP in 4% (6/167) of melanomas (p.Ser722Phe). The likelihood of this occurrence is ~ 5 X 10-20, suggesting this mutation is functionally significant. TRRAP is part of a protein complex that has histone acetyltransferase activity and is an essential cofactor for the oncogenic transcription factors c-Myc and E1A/E2F. Functional experiments show that the p.Ser722Phe mutation is transforming and that mutant TRRAP is required for survival of melanoma cells; insights into how mutant TRRAP contributes to carcinogenesis is an important area for future research. The identification of TRRAP mutations adds melanoma to the long list of cancers harboring mutations in proteins with histone modifying activity.

Chimpanzees are selfish, but children are kind

I pose an age-old question: what is it that makes us human? I think it depends who you ask. Ask a cognitive neuroscientist and they may say it’s our theory of mind, which is a fancy way of saying humans have empathy. Ask an evolutionary biologist and they will likely point out all the morphological traits that distinguish us from other primates, such as the large size of our cranial vaults or our opposable thumbs. Ask a psychologist and they may cite our conscience or our ability to use symbolism. But no matter who you ask, most would likely agree that our capacity for sharing resources and the social rules that regulate sharing are specific to human culture.

In a new study published in Nature,

Michael Tomasello and colleagues report a series of experiments in 2-3-year-old children and chimpanzees and conclude that 3-year-old children are more likely to equitably divide resources gained by collaborative activities compared to a non-collaborative situation. They further conclude that collaborative activity does not seem to influence sharing in chimpanzees.

It is known that 3- to 4-year-old children typically divide resources unequally and are more likely to maintain possession of any resources for themselves than share with others. However, as children approach five to seven years of age, they start to share resources more equitably. Tomasello and colleagues hypothesized that children might share a resource more equitably at an earlier age if they had to work collaboratively to attain it, compared to when resources are provided by adults (as resources usually are when children are 3-4 years old).

In this study, pairs of age-matched children were put in a room by themselves (after a demonstration phase) and exposed to a collaboration, no-work, or parallel-work condition. In the collaboration scenario, the children faced a rope attached to an enclosed apparatus. Pulling on the rope together would bring toys in the enclosed apparatus toward them, and the ‘lucky’ child would gain three toys, while the ‘unlucky’ child would gain one toy. In the no-work condition, the children entered the room with the toys already positioned in the end-state. The ‘lucky’ child shared the extra toy with the ‘unlucky’ child more often in the collaboration condition compared to the no-work condition. To control for the possibility that sharing was influenced by the fact that the collaboration condition required work, the authors also ran a parallel-work condition. Here, each child had to pull on their own rope to attain the toys. The authors still observed that the ‘lucky’ child shared toys with the ‘unlucky’ child more often in the collaboration condition compared to either the no-work or parallel-work conditions.

Next, the authors hypothesized that collaboration would not influence resource sharing in chimpanzees. They placed pairs of chimpanzees on opposing sides of an apparatus that required simultaneous pulling of a rope to move grapes to a see-saw that was accessible to both chimpanzees. The ‘lucky’ chimp attained 2 grapes and a chance to take the other grape, while the ‘unlucky’ animal received 1 grape and also had a chance to take the other grape. In the first study, the ‘unlucky’ chimp tipped the see-saw and took the other grape for itself in 63% of trials. The ‘lucky’ chimp never actively tipped the seesaw toward the unlucky animal. Next, the authors tried to encourage the ‘lucky’ chimp to share by disabling the seesaw to the ‘unlucky’ chimp. They noted that the ‘lucky’ chimp tipped the food to itself in 98% of cases, and found no difference in sharing in collaborative vs. control conditions.

This study demonstrates that humans as young as 2- to 3-years-old are able to recognize rewards attained through collaborative efforts and demonstrate a sense of “distributive justice.” Since chimpanzees do not require collaboration for acquisition of resources in the wild, the authors suggest that chimpanzees have not, as a species, developed this sense. This study provides evidence that a sense of “distributive justice” distinguishes humans from other primates and reinforces the notion that collaborative efforts played an important role in human evolution. And so I end with another question. Are humans inherently kind or selfish? I don’t think we know the answer to that question, but this study implies that for humans, evolution has favored the kind.

Archon Genomics X Prize in brief

As I understand it, the contest is technology-driven to get 100 genomes in 30 days. The benefits to us in the genetics community is 100 excellent genomes sequenced many times over, from 105 year-olds who act as supercontrols for age-related diseases. We also get a software contest so we get better browsers. I think if the 100 centenarians can be taken from 10 different projects, then we can leverage the projects via the shared controls and get more data sharing and trust

Setting standards for medical genome sequencing

As described in the editorial “Towards a medical grade human genome sequence”, we are pleased to collaborating with the $10 Million Archon Genomics X PRIZE (AGXP) to develop a set of standards and procedures to help with judging the prize contest. Your comments are needed but should not be limited to this aim alone because the contest is just the beginning of this standard-setting project. For examples, it is likely that the very kinds of genome variation browser and related software that are needed to evaluate the sequencing effort are the spin-off product the community really needs to make medical genomics a reality.

The draft validation protocol is a collective assembly of techniques designed to test the quality and accuracy of 100 whole human genome sequences resulting from the AGXP competition. The purpose of this article is to enlist constructive criticism from the genomic and genetic community on the outlined approaches.

Alexander Zaranek et al. have in response posted an intriguing set of suggestions to simplify, improve and reduce costs of validating the AGXP. We encourage readers to examine that article and help us decide whether the suggestions should be substituted or incorporated into the Protocol.

Thomas Perls and Nir Barzilai have now made the exciting suggestion that the AGXP use the genomes of 100 survivors of age-related diseases (over 105 years old) as the sample set for the contest. They list a number of reasons why such a set of genomes would add considerable scientific knowledge as opposed to 100 random genomes.

Comments can be submitted via Nature Precedings, or can be sent directly to the corresponding author of the validation protocol: Larry Kedes kedes[at] Alternatively, you may prefer to email this Nature Genetics blog at freeassociation[at] In the latter case, please let the moderator (Myles Axton) know how you wish your comments to be posted here. They can be either attributed or anonymous, provided you let us know your email address and real name.

Genetics of treats kindly treated

Dear Myles:

I truly enjoyed reading the following statements in your editorial (43:page 85) 

First…“for the parallel study of mechanisms of evolution, both the mechanisms of speciation and those operating in adaptation at the population level in the course of domestication history and agricultural improvement.” 

Second. “…we remain at heart a journal of genetic variation.” 

I admire your commitment to the above principles. 

Darwin was inspired by domestication of plants and animals by farmers and animal fanciers, and he asked how variation was molded by nature. 

The great triumvirate – Fisher, Haldane and Wright were not only inspired by Darwin, but also intimately associated with plant and animal breeding. That is why their work is so practical, meaningful, endearing and enduring. Perhaps it may not be an exaggeration if I mention that to a large extent, we are yet to transcend the limits (should we call it the Fisher-Haldane-Wright limit?) set by these visionaries.

 The later ones such as Kimura, Ohta, Nei, Mather, McClintock, Beadle to name a few, were all trained in plant breeding. Similarly, Lush, Robertson, Cockerham, Henderson et al. were animal breeders. Examples such as the origin of cultivated wheat, cotton and sunflower are exhilarating. Indeed, these discoveries absolutely dwarf the present craze in structural/copy number variation – Please don’t mention this to Evan Eichler or Jim Lupski! 

Single genes could transform and even destroy agriculture and livestock. I am sure humans are no different from these general phenomena. These views must be addressed in our own times. 

It is a pleasure to witness this wonderful progress and to see a journal committed to advance these fundamental principles of biology. 

Sincerely yours

- Raju Govindaraju

Cropping and editing: crediting

Dr Raju Govindaraju takes me to task for skipping from 1906 to 1919 without crediting the advances in genetics from the field of plant breeding. It is a fascinating history deserving of books rather than just a paragraph in an editorial. I thank him for his comments:

“I congratulate you on your Editorial titled, “In praise of maize.” (Nature Genetics 42: 1031). It is heartening to note that Nature Genetics is finally taking a look at other organisms and questions (in addition to human GWAS!) that also throw light on human genetics and health.

I would like to make the following remarks on your editorial:

1. Although the title rhymes well…perhaps “In praise of crops,” or plants (Arabidopsis or yeast or other plants are not crops) or something similar would have been more appropriate.

2. The sentence “The discipline (i.e., the science of genetics – I suppose) ….current genetic methods…” is misleading as it glosses over many important points, and gives the impression all of that happened only after 1919.

a. The big intellectual boost for genetics came in the 1910’s using plants are: Inbreeding and hybrid corn (East, Shull, Jones); pure line theory, gene, genotype, phenotype (Beans; Wilhelm Johanssen); size inheritance (Tobacco; East); size and shape in plants (Emerson); wheat color (Nilsson Ehle) etc.

Fisher’s monumental work of 1918 was the predecessor of his 1925 work. The latter introduced “how to do experiments” with replication, randomization and local control. Harvard awarded him an honorary degree citing this work. In other words, “the big boost” happened prior to 1919. Only “Modern Synthesis” happened following Fisher’s 1925 work in the early 30’s (Fisher, Haldane and Wright and others later)c. Rothamstead’s focus on maize research was minimal. They focused largely on the application of fertilizers and crop yields.

Thank you for your interest.


D.R. Govindaraju"

Update on post-GWAS functional standards paper

It is interesting to see how new views are building up as more authors join the initiative to develop the draft standards paper. It is also fascinating to see them experiment with new tools and services for a diverse author group. There is now an active discussion page for authors and referees of the standards paper at Wikigenes. A collection of relevant supporting papers has been assembled on Mendeley. It would be great if the corresponding authors have time to submit a revised draft to Nature Precedings before the authoring deadline so that we can see how the reference copy has evolved.

Wikigenes joins our experiment in collaborative authoring of standards

Now there is another way to make your contribution to the standards document “Principles for the post-GWAS functional characterisation of risk loci” described in the previous post, automatically becoming a co-author in the process. Wikigenes creator Robert Hoffmann has set up a splash page, introducing the paper and a link to the editable version at Wikigenes.

We do not anticipate problems in assigning authorship and author roles since it is obvious whether you have contributed conceptual input or corrected punctuation. Corresponding author Ian Mills and editor Myles Axton will check and if necessary edit author names and roles for the final version to be submitted for publication after December 20th 2010.

To follow changes of the article you can subscribe to the following feed

If you have problems with the feed please address your queries to Wikigenes, not to Free Association blog, bearing in mind this is an experiment.