At the last evening keynote address for the MOHG, Jeremy Berg, the head of the National Institute of General Medical Sciences presented some early data coming out of the NIH’s new peer review policies instituted last year. As part of the new system, members of the review committees for grants are instructed to rate individual applications on a scale of 1 to 9 for the quality of five different categories: significance, investigator, innovation, approach, and environment (meaning the institutional support available to the investigator). Although these criteria are not calculated into the overall score upon which research decisions are made, Berg says they can serve as a parallel indicator of what the review panels seem to value the most. Based on 360 grants to the NIGMS as of October 2009, the criteria most predictive of getting a high overall score were, approach (.74), significance (.63), innovation (.54), investigator (.49), and environment (.37). Grant administrators, says Berg, use the data to look closer at, for example, why some highly innovative grants aren’t being awarded high overall scores and to make sure that the emphasis on approach doesn’t mean that study sections are favouring projects that look reasonable and therefore possibly ‘safe.’ It provides, Berg says, another way of looking at our portfolio.

For information on other scientific indicators and how they’re being used, check out our metrics special.

At the MOHB today, Alaine Keebaugh of Emory University presented work that helps to explain a 20-year old puzzle in mouse and human genetics. In the late 1980s Mario Cappechi and Oliver Smithies disrupted the first gene in mouse embryonic stem cells, a mouse homologue of the human gene, HPRT1. The gene was a useful target because it was on the X chromosome, meaning that they only had to knock out one copy in a male embryonic cell line to completely abolish production of the protein. And they had a way to test that the protein had been eradicated. The knockout mice that resulted from this work transformed genetics and earned Cappechi and Smithies a Nobel Prize. But surprisingly, aside from the verifiable lack of the protein, the mouse was unremarkable, not very different from wild type.

In humans, on the other hand, disruption of HPRT1 results in a devastating disorder known as Lesch-Nyhan disease which has confers gouty arthritis, atrophy of the testicles, developmental delay – they never learn to walk – and they tend to be somewhat aggressive. But most puzzling is the severe self-injurious behaviours, including biting and chewing off of fingers, lips—pretty much anything they can get their mouths on. They don’t want to hurt themselves and many have to be restrained for much of the time. Some can even warn their caregivers that an uncontrollable urge to harm themselves is coming. If you haven’t read a heart-wrenching story from The New Yorker on the topic, (subscription required) I recommend checking it out.

Keebaugh was looking to develop a better mouse model of the disease, so she searched for other similar genes in vertebrate genomes and came across the similarly tongue twisting PRTFDC1, which appears to be an ancestral duplicate of the gene that is active in humans but appears to have been inactivated in the mouse lineage.

Suspecting that the difference in activation might somehow account for why knockout mice don’t show much of a phenotype, she engineered a transgenic line of mouse to express the human version of PRTFDC1. When she crossed this with HPRT1 knockouts she found that the offspring males, containing both the active PRTFCD1 and the inactive HPRT1 display some behaviours similar to Lesch-Nyham males. They’re more aggressive, and while they don’t have the neuromuscular defects, when given amphetamine (a standard way to study stereotypical behaviour, apparently) she observed them doing something she’d never seen mice do before, standing on their haunches and nibbling away at their fingernails.

At the MOHB this morning, Pamela Sklar of Massachusetts General Hospital presented data from the Psychiatric Genome-wide Association Study Consortium (PGC). One of its projects on bipolar disorder looked at the genomes of more than 7,000 cases of the disorder against 10,000 or so matched controls trying to find differences that correlate with an increased risk for the disease. What they’ve found was not different from what many genome-wide association studies have found. Four strongly associated genetic regions popped up in their study, but the individual amount of risk for the disorder that they contribute is quite low. For this reason and because many of the variants they find associated with bipolar and other psychiatric disorders aren’t necessarily gene mutations (many have been genetic duplications and deletions), says Sklar, it’s been hard to convince biologists with experience in model systems to delve in and look at how the gene variants might be contributing.

Nevertheless, working with a collaborator with expertise in fly genetics, Sklar presented data on the function of three genes that might be linked to bipolar disorder. They manipulated versions of the genes in Drosophila melanogaster and observed the formation of synapses between neuron and muscle. At these synapses the long neuron cells form little bulbous protrusions known as boutons and these form in well documented ways in Drosophila larvae. When the researchers knocked out expression of their suspect genes in the flies, the neurons formed some normal looking boutons but also some “ghost” boutons that appeared to be immature non-functioning synapses. The results will require more follow up, but are interesting. The GWAS and genome sequencing studies that Sklar and others are doing are producing “real risk genes in schizophrenia and bipolar disorder,” she said to an audience filled with researcher who work primarily on model systems. “They need people like you to study them.”

At the MOHB meeting this morning the topic was sex, specifically that topic sure to stir up the hornets, the differences between sexes. Nirao Shah of UCSF and Melissa Hines from the University of Cambridge talked about how hormone levels might be responsible for shaping brain differences. But Eric Vilain, who studies intersex individuals at UCLA offered some surprising takes on the differentiation in the brain as it may be shaped by epigenetics, that is environmentally influenced changes that don’t affect the sequence of the genome but alter its expression. Specifically he was looking at methylation of DNA, which marks active and inactive genes in a few dozen pairs of maternal twins discordant for sexual orientation. Tongue firmly in cheek, he referred to the field of study as “epigayomics” in his slide presentation. But the results were negative, in 34 twin pairs they found very little difference in the way genomes were methylated between gay and straight males. His group’s research confirms that as twins age their epigenetic profiles diverge more and more, but the maximum difference was very low, he said, calling the methylation patterns “exquisitely similar.”

The first set of talks at the MOHB in Boston yesterday evening was on personal genomics. Speakers including the Broad Institute’s David Altshuler and Leonid Kruglyak of Princeton vigorously defended the reputation of genome wide association studies (GWAS), which have come under some attack in the New York Times and other media outlets, recently. GWAS compare genomic markers in hundreds to thousands of individuals in order to find areas of the genome that associate with risk for common traits and diseases. They’ve produced hundreds of associations, but frustratingly for some, the genetic regions that have been fingered account for only a small percentage of the inherited risk for disease. Nature weighed in on this ‘missing heritability’ concept in the past (see here and here).

Altshuler takes umbrage with some, such as Mary Claire King at the University of Washington who have posited that very rare variants – unlikely to be found by GWAS – are where much of the heritability is hiding. Altshuler says he thinks that with projects like the 1,000 genomes project (which now plans to sequence upwards of 2,500 human genomes to various degrees of completion) a lot more human variation will be found and will begin to fill in more of the gaps in heritability. As costs of sequencing come down, he says, “We won’t have these debates about rare versus common variants.”

Capturing a real sense of human diversity in genomes will be important says Carlos Bustamante, now at Stanford University, and that should be done by sequencing more humans from diverse backgrounds (rather than just European backgrounds which have dominated the sequencer queues to date). His talk presented some new work looking closely at the sequence of an African American woman and a Mexican individual. Both individuals have a genome that is admixed, that is it contains a jumble of genetic elements from European and African origin in the case of the African American and Native American and European. Using computational tools, Bustamante’s group was able to deconvolute the ancestry of the genome in an easy to see way showing where stretches of DNA on the individuals’ genomes became mixed. The longer the stretches go without interruption, the more recent the admixture, says Bustamante, and they were roughly able to calculate the point at which cultures came together – lining up roughly with the Spanish exploration of the Americas (1470-1570) for the Mexican individual and the heyday of the middle passage (1690-1750) for the African American. It was a pleasing kind of “parlor trick” that these dates meshed so well with what we assume about history, he says. “We need continued sequencing of very diverse human genomes,” says Bustamante. It’s the only way we’ll be able to know what we’ve been missing.

In Boston this weekend, researchers from diverse fields and backgrounds have converged to discuss how traditional model organisms like yeast, flies and the worm C. elegans, promise to contribute to the understanding and hopefully the treatment of human disease. It’s the third biennial meeting called Genetics 2010: Model Organisms to Human Biology (which has the inexplicably pleasing acronym, MOHB). Scott Hawley of the Stowers Institute and current president of the Genetics Society of America, which organized the meeting, is a fly researcher. In his opening remarks yesterday evening, he noted that the divide between biologists studying human biology and those studying model organisms is often too great. “We want to reach out and have more contact with people doing other things,” he said. With a line up of talks on everything from personal genomics to sex determination neurogenetics and infectious disease, it promises to deliver that kind of contact.

Cancer genomes have been a hot topic at this year’s AACR. I stopped in to see a session hosted by Elaine Mardis, Washington University’s genome maven whose been an author on most of the big cancer genome papers to date. In the session, we heard from Todd Golub of the Broad Institute, who gave preliminary results on the multiple myeloma genome, which hasn’t yet been published.

It looked like it has produced several interesting new potential cancer genes to look into. Though I won’t go into too much detail, here are some of the basic stats. They looked at cancers from 38 individuals sequencing both cancer cells and normal cells. They fully sequenced 23 of the individuals, and they did what’s called whole exome sequencing for 16 (in which they just sequence protein coding regions). One patient was sequenced by both methods.

The data produced a big list of mutations, but researchers have learned tricks for paring down such lists to find the so-called ‘drivers’ of cancer. By, for example, looking for mutations that appear frequently in cancer cells from different individuals. They came up with a short list of a dozen leading candidates. Four have been well characterized. Among the others, Golub found genes involved in regulating translation, the process by which RNA is made into protein, and even a gene implicated in susceptibility to Parkinson disease.

More research presented today at AACR’s 101st annual meeting shed some light on the mind blowing complexity of cancer. At this morning’s plenary sessions, Alan Balmain of the University of California San Francisco showed how the simple model of cancer initiation leading to progression and metastasis was a vast oversimplification. Cancer cells, he says require help from otherwise normal stromal cells, blood vessels, and inflammatory cells. And while much of the research presented at this meeting has been about cataloguing mutations that are gained in cancer, he’s been trying to better understand the underlying genetic background that plays a role in intrinsic susceptibility to cancers.

It’s hard to pull such genes out from studies of humans, so he crossed two strains of mouse, one that is susceptible to cancer and one that is relatively resistant, essentially creating a heterogeneous population of offspring with variable susceptibility to cancer. On these mice he did gene expression analysis for different skin samples and tumor samples. This analytical approach helps to uncover genes that are working in concert to influence cancer susceptibility, thus exposing deeper networks of genes at play in the process that can have interlinked function. Mice that were susceptible to tumors, for example were enriched for expression of genes involved in determining an epidermal skin cell fate, as opposed to a sebaceous or follicle fate. Genes involved in mitosis were upregulated, wound healing genes were upregulated, and genes for reigning in the inflammatory response were upregulated.

Balmain went a bit more into depth on the inflammatory genes however. Generally inflammation is generally associated with increased cancer susceptibility, but anti-inflammatory drugs have different effects on the development of skin tumors. Sometimes increasing tumor spread and other times preventing against it. Balmain’s systems biology approach has indicated a number of genes related to controlling inflammation and shows how they could be related to cancer susceptibility in his mouse population. See a paper on it from last year, here. But it’s complex relationship. It’s times like this that I consider the Thermos: it keeps things hot and it keeps things cold. Whenever something seems to have an important job in biology like preventing cancer, it almost always does exactly the opposite with a subtle shift in context.

Arul Chinnaiyan kicked off the day for AACR’s 101st annual meeting in DC by talking about cancer genomes. He gave a roundup of some of the major genomes published to date, many of them in Nature. He even showed a brilliant screenshot of Heidi Ledford’s April 15 feature on the topic. Bert Vogelstein of Johns Hopkins followed up with a talk that seemed too good to be true, asserting that thanks to decades of research, cancer is essentially a known entity. He’s been comparing the genomic landscape of nearly 100 human cancer genomes that have been sequenced to date and other data to come up with 3142 genes that are mutated regularly. He applies a simple criteria to distinguish which of these three thousand genes likely suppress tumor formation as their usual function (a function which a mutation disrupts), and those that through mutation become more active and cause cancer (so-called oncogenes).

Continue reading →

Today there was a lot of buzz surrounding the release of results from the BATTLE trial (Biomarker-integrated Approaches of Targeted Therapy for Lung cancer Elimination), sponsored by the US Department of Defense. This trial, started in 2006 attempted to group patients by predominant biologic features of their cancer, including those that can be characterized by genetic changes in EGFR, KRas, RXR/CyclinD1 or VEGF (all predominant defining molecular signatures lung cancer) and see if they could match these patients with the optimal treatments choosing from four treatment regimens.

Continue reading →