We’ve moved!

Our blog has a new home and a new look! Find us at: http://blog.nextbio.com/

Please update your subscriptions and feeds to make sure you receive all our latest posts.


Crowdsourcing solutions for life

From the lab to scientific publishing to clinical solutions, the Open Access movement is changing the way science advances

(Guest post by Joseph Jackson)

The 2nd Open Science Summit is the weekend of October 22-23, 2011 at the Computer History Museum in Mountain View, just before the 2011 Open Access (OA) week, which runs October 24th-28th.  Open Science evokes different associations for different people, depending on which part of the scientific process they most regularly engage with.  One critical component focuses on access to scientific literature.  The OA movement has made great strides in the last ten years with the creation and maturation of journals and publishers like PLoS and Biomed Central.

But the most critical shifts toward Open Science arguably are happening in the life sciences.  The technological revolution underway in next generation sequencing is enabling, but also requiring bold new collaborative approaches to manage increasing complexity and accelerate the translation of scientific discoveries into desperately needed therapies.

The success of Open Biology depends on the formation of an ecosystem where companies, academics, public sector researchers, and citizen scientists can all participate and create shared value.  NextBio itself is already taking steps towards supporting the creation of this new infrastructure with NextBio Public.  Another significant development is the announcement by DNAnexus that it is hosting the Short Read Archive (SRA), a critically important public data resource whose funding was being phased out, leaving its future in jeopardy.

Increasingly impressive examples demonstrate the success of Open Science in solving formerly intractable problems.  Take for example Foldit, an online game that allows “citizen scientists” to participate in modeling protein folding.  In September it was announced that a team using Foldit solved the structure of an enzyme critical to the reproduction of the HIV virus, an insight which may yield new therapeutic targets, and which had eluded scientists for the past ten years.  This is only one example of harnessing the power of crowdsourcing for innovation.  Tomasz Sablinski, a long time veteran of the pharmaceutical industry and formerly of Novartis, has launched Transparency Life Sciences to apply the “wisdom of the crowd” to the most expensive part of the entire biomedical research process, the design and implementation of clinical trials. Another project to watch is John Wilbanks’  Consent to Research initiative, which seeks to build extensible standards for patients’ to participate in research in a way that ensures their results are available to help everyone.

What does this mean for the future of privacy and informed consent?  We will examine these questions and more at the Summit when we look at the future of Open Medicine.  While many models are being tried, it is clear that the traditional biomedical establishment, which seeks to ensure safety, quality, and ethical treatment of human subjects through mechanisms such as Institutional Review Boards (IRBs), is struggling to adapt to a patient powered paradigm.

The future of Open Science is brighter than ever, but it needs your involvement to fulfill its great potential.  Join us at the Summit or watch the live stream at  http://fora.tv/live/open_science/open_science_summit_2011.

Guest Blogger Joseph Jackson is the founder of the Open Science Summit, co-founder of BioCurious, a community lab effort in the Bay Area and the CEO of LavaAmp, Inc. Follow @OpenScienceSum on Twitter, and check back with us for more blog posts about talks at #OSS2011.

SIRT2 finds a place in the family

Sirtuin 2 may suppress tumor formation in mice and humans

Sirtuins are a family of proteins that have been implicated in processes ranging from aging and tumor formation to obesity and cerebral ischemia. Seven sirtuin proteins are known to exist in mammals, two of which (SIRT1 and SIRT3), are known tumor suppressor genes. The function of sirtuin 2 (SIRT2) remained unclear however, until recent research from scientists at the National Institute of Health, Vanderbilt University and the University of Texas, Dallas discovered that SIRT2 could play an important role in preventing cancer. In a paper published in Cancer Cell this week, the researchers describe the role of SIRT2 in maintaining genomic stability in cells, a function critical in aging and cancer-related processes.

The scientists began their experiments by eliminating expression of the SIRT2 gene in mice. The absence of SIRT2 had no apparent effects on embryonic and early development in the mice, nor did their organs show any obvious abnormalities. However, cultured mouse embryonic fibroblast cells (MEFs) from the mutants grew more slowly than normal (wild-type) fibroblasts, and stopped growing earlier, implying that SIRT2 is important for cell proliferation. The mutant cells also had more errors in chromosome segregation during cell division: 35% of mutant cells had more or less chromosomes than normal (less than 5% of wild-type cells had similar defects).

Could these errors in cell division affect cancer risk? Observing the SIRT2 mice over a 2-year period, the researchers found that 60% of SIRT2 mutant mice developed some form of cancer by 24 months. Unusually, tumor formation in these mice showed a gender-specific pattern. Female mice developed breast cancers, whereas males developed tumors in multiple organs, including the lungs, pancreas, stomach and prostate.

Previous research has suggested that SIRT2, a histone deacetylase, is essential during mitosis and is associated with forming chromosomal structures for normal cell division. Exploring the molecular basis of this function through several experiments, this new research reveals interactions of SIRT2 with molecules critical to normal mitosis in cells, such as Aurora proteins and the anaphase promoting complex APC/C. Cells lacking SIRT2 end up with many abnormalities in cell division, including centrosome amplification, metaphase arrest and cell death during mitosis.

On chr. 19, the SIRT2 gene is located near the centromere

The authors also extend their experiments to human cancers, finding that breast and hepatic tumor samples expressed the SIRT2 protein at levels about 1.5 -2.5 times less than (corresponding) normal tissues. Examining existing data in Oncomine, they found similar trends in other human cancers, like renal carcinomas, glioblastoma and prostate carcinomas. Though there’s no obvious explanation for the gender-specificity of the tumors that mice developed, this research strongly supports further studies on SIRT2 function in cancers.

Exploring the SIRT2 gene in NextBio’s Disease Atlas, we found nearly 600 studies linking SIRT2 disruptions to several cancer types. We restricted our search to human studies of gene expression differences between diseased and normal tissues (currently 120 studies). We selected 23 biosets for meta-analysis, and found SIRT2 expression was altered at least 1.5 fold (relative to normal tissue) in all 23. Most primary cancers showed a decrease in SIRT2 expression consistent with the observations described in this study. However, we found that some neural tumors appear to show an increase, rather than decrease, in SIRT2 levels relative to normal tissues. The fold-change in SIRT2 expression also seems to be higher in metastatic breast tumors relative to primary tumors. Prostate cancer also shows a similar trend, with more advanced stages and metastatic tumors showing an increase in SIRT2 expression.

Fold-change in SIRT2 gene expression in cancer samples relative to normal tissues. *denotes average of at least 2 datasets, SD < 0.05

NextBio analysis of public genomic data supports the conclusions of this study with evidence from experiments on cell lines, tissues and a variety of clinical samples. Analysis of somatic mutations or differences in DNA methylation in the SIRT2 promoter region could offer further insight into the significance of SIRT2 to specific sub-types of human cancers.

Intern Life at NextBio

Interns at NextBio learn to set the stage for a unique kind of data exploration

At NextBio, genomic data snakes through the hands of scientific teams and the automated pipelines they design, connecting people intellectually and socially. Each department is responsible for their own piece of the NextBio puzzle as well as helping new team members cultivate their skills.

The data curation team at NextBio mimics the work of a heart, channeling in the public genomic data that’s essential to the NextBio platform. Beatrice Chiu, who graduated from the Molecular and Cell biology program at UC Berkeley, began her journey at NextBio as a Web Product intern, conducting usability tests to optimize the NextBio user interface. She switched over to the curation team earlier this year to help with a large scale GWAS (genome-wide association study) tagging project. As Beatrice explains, “All studies in the NextBio database usually have a minimum of a biodesign tag, like disease vs. normal, response to a drug, etc. and then a more specific phenotype tag, say for a disease. GWAS studies can also get classified using case-control or other association tags.”

“This helps users filter data in so many different ways, rather than just looking for studies on a particular disease or method,” she adds. Beatrice explains that working in both the Curation and Product teams “helps her understand different aspects of the NextBio correlation engine fully.”

Data imported by curation is then analyzed by the Bio-Computing team, where Meenakshi Mali, Computer Science graduate from San Jose State University, interns. Meenakshi first found out about NextBio through a friend, and was “amazed at the impact of NextBio’s work in this field and how the data processing here is revolutionizing the way researchers work.” Specifically, “researchers can use this information to find correlated studies and make hypotheses based on this data,” explains Meenakshi. The bio-computing team creates and uses different algorithms to process curated data and integrate it into different features on the NextBio site. When searching for a specific study, users can view the relative score of different parameters to construct their own research experiments. This technology is part of what inspired Meenakshi to bring her scripting skills to NextBio. Meenakshi points out that “dealing with the basic scripts, that make a product [like NextBio]” is one of her favorite aspects of working here. On a typical day she conducts quality analysis to improve the algorithms used to analyze curated data, making sure that improvements in the statistical analysis are accurate and error-free.

Qing Zhang, Engineering intern and graduate from the University of Wisconsin Milwaukee, also knows a thing or two about coding. She assists in building the search engine that powers NextBio’s databases by designing data mining programs that are as accurate, efficient, and fast as possible. After all, crafting high-precision software is essential to maintain a database that integrates (and indexes) users to index over 20 million PubMed documents. Improving processing time for each document by merely one millisecond can reduce the time it takes to process a document by five hours. The incredible amount of information and genetic data that high-throughput technology has made available to the public keeps the entire NextBio team on their toes.

Nonetheless, all three agree, the continual teamwork and collaboration adds to the rewarding and enriching experience. For Beatrice, her favorite part of working at NextBio is the people: “Working here is truly a group effort,” she says. Qing notes that it is this effort that has produced “a sophisticated system used by several thousands of people.” So what does the future hold for this group of NextBio interns? Qing loves the fast pace and excitement of industry, and wants to continue perfecting the world of data mining. Meenakshi plans to delve deeper into the statistical analysis side of bio-computing, embracing the human calculator. And Beatrice? Well, she would like to be the boss…but all joking aside, she plans to continue her journey in data curation, populating the NextBio database one study at a time.

Nicer with Dicer? Not always.

Increases in Dicer gene expression improve some cancers and worsen others

Some of the tiniest players on the field, microRNAs (miRNAs) are short strands of RNA that work to control gene expression in several pathways. Though they don’t encode any proteins, miRNAs regulate genes involved in embryonic growth, cell differentiation, angiogenesis and other cellular processes. They have also gained steady prominence in cancer research, with several studies connecting abnormal miRNA regulation to cancer progression and metastasis in cell lines, animal models and samples from patients.

The potential diagnostic and therapeutic value of miRNAs has also turned attention to the molecules involved in making miRNAs in cells, particularly the enzyme Dicer. A recent study by Zhihai Ma and colleagues in PLoS One reports that higher Dicer expression correlates to more advanced stages of cutaneous melanoma, characterized by increased metastatic potential, tumor mitotic index, and other stages (as defined by the American Joint Committee on Cancer) .

Other studies have identified specific miRNAs linked to enhanced proliferation, metastatic potential, post-recurrence survival of patients, etc. but this is the first research examining the correlation of Dicer levels, rather than miRNA profiles, to melanoma progression.

Source: Wikipedia

Comparing amounts of the Dicer protein in over 400 samples of different kinds of skin cancers, Ma et al. found that Dicer levels were significantly greater in cutaneous melanomas than in carcinomas, sarcomas or benign melanocytic nevi. Testing the location and amounts of the Dicer protein inside cancer cells, the authors discovered that normal skin keratinocytes express Dicer at consistently low levels. Different kinds of carcinomas and sarcomas also express low levels of the protein, but primary cutaneous and metastatic melanoma cells showed high levels of Dicer. Amongst melanocytic tumors, nevus cells and mucosal melanomas had lower levels of the protein than cutaneous, acrolentiginous and metastatic melanomas.

The authors support these results by mining existing research on the RNA expression profiles of genes involved in miRNA synthesis and processing in cells. Using NextBio, they pooled data from two studies on expression patterns of all the enzymes involved in miRNA synthesis and processing in cells, combining data from over 25,000 genes spanning 20 disease groups, including normal skin and several cutaneous cancers. Here, Dicer ranked among the top 20% of all significant genes, and was expressed 2.5 fold higher in cutaneous melanomas than in basal cell carcinomas.

Overall, their results show that an increase in Dicer expression strongly correlates to metastatic and more severe forms of cutaneous melanoma rather than normal skin or benign nevi. This research adds to a growing body of evidence associating Dicer to cancer in patient tumor samples.

Dicer is already considered a prominent, albeit confusing, player in prostate, lung, breast, ovarian and several other cancer types. In some cancers, like prostate adenocarcinomas, an increase in Dicer expression is linked to poor prognosis, whereas in others, such as lung and ovarian carcinomas, lower Dicer expression is linked to more severe disease. These contradictory observations are particularly important since most of these studies conclude that Dicer expression is a strong candidate biomarker or potential drug target.

Experiments in mouse models offer a potential explanation of Dicer mechanisms in cancer. Two recent studies created mouse mutants of Dicer to test the effects of the mutation on retinoblastoma, a kind of eye cancer. Both studies, one by Kumar et al. and the other by Lambertz et al., identify a dosage-dependent effect of Dicer. In their experiments, knocking out one copy of the Dicer gene enhanced tumor formation in the mice, but knocking out both copies led to suppression of tumor formation. Their results suggest that partial loss of Dicer function favors tumor formation, while completely eliminating Dicer expression suppresses tumor formation, at least in retinoblastomas. Though these studies suggest one explanation, several other possibilities exist (tissue-specific effects being the most obvious) and need to be verified experimentally.

So far, research on the role of Dicer in human cancers has largely focused on establishing correlations between Dicer levels and cancer prognosis. Studies like the cutaneous melanoma research described here add to this understanding of Dicer levels in different kinds of cancers and increase the possibility of Dicer being used as a potential drug target, biomarker or diagnostic tool. In this context, a better understanding of the mechanisms by which Dicer and its product miRNAs exert their effects in human cancers is especially critical.

Announcing: NextBio Public

If you’re a user of NextBio Basic, you’re probably aware that NextBio lets you access thousands of gene expression studies from sources such as GEO. NextBio has done the heavy lifting of importing, normalizing, and curating publicly available genomic data so that biologists can ask and answer biological questions completely in silico—even with little computational expertise.

Which makes our recent updates very exciting: We’ve made data from GWAS, CNV studies, and more freely available* through the NextBio Public site. These numbers update on a near-daily basis. Because our curators tag each imported study using a controlled vocabulary and custom disease ontology, you’ll easily be able to perform meta-analyses of multiple studies across different diseases through a simple, web-based interface.

Correlating genomic data from thousands of studies linked to different diseases, compounds, tissues, cells and more is complicated. We’re constantly trying to make the job a little easier, which is why we’ve also given the NextBio Public user interface a major overhaul.

NextBio Public brings you the same user-friendly “Vertical Apps” interface that NextBio Enterprise subscribers use. We’ve included a familiar element by borrowing the term ‘apps’ for our applications, the centerpiece of our new interface. Think of each app as answering a specific biological question. Come explore our new look and the abundance of new data, and don’t forget to come back for tours and examples of the many ways NextBio can help you generate or validate hypotheses, conduct in silico experiments, and much more!

*NextBio Public is freely available to users from academic, government and non-profit research institutions. If you have a Basic account, you get an automatic upgrade to NextBio Public- just don’t forget to update your e-mail address!

Quiz winner!

If you haven’t already, check out our first blog quiz. Though the contest is closed, you’re still welcome to guess the answers! For the curious, the correct answers are included here.

Though we’d announced multiple prizes, only one entrant got all the right answers: Congratulations (once again) Dave Schlesinger!

The answers:

1. The Translational Genomics Research Institute

2. MCU

3. Paroxetine

4. Sepandar Kamvar

5. No one knows

6. All of them