Tuesday, January 6, 2015

Publish or Perish: an old system adapting to the digital era

    By Annie Chen and Michael Allegrezza
      
           When scientific publishing was developed in the 19th century, it was designed to overcome barriers that prevented scientists from disseminating their research findings efficiently. It was not feasible for scientists to arrange for typesetting, peer review, printing, and shipping of their results to every researcher in their field. As payment for these services offered by publishers, the researchers would transfer the exclusive copyrights for this material to the publisher, who would then charge subscribers access fees. To limit the printing costs associated with this system, journals only published articles with the most significant findings. Now, nearly 200 years later, we have computers, word processors, and the Internet. Information sharing has become easier than ever before, and it is nearly instantaneous. But the prevailing model of subscription-based publishing remains tethered to its pre-digital origins, and for the most part these publishers have used the Internet within this model, rather than as a tool to create a new and better system for sharing research.

Figure 1. Trend lines show an annual increase of 6.7%
for serials expenditures vs 2.9% for the Consumer Price
Index over the period 1986-2010, relative to 1986 prices.
In theory, digitization should have decreased costs of communicating science: (authors can perform many of the typesetting functions, articles can be uploaded online instead of printed and shipped, etc. In practice, however, digitization has actually increased the price of journals. Statistics from the Association of Research Libraries show that the amount spent on serials increased 6.7% per year between 1986 and 2011, while inflation as measured by the US Consumer Prices Index only rose 2.9% per year over the same period (Figure 1).1 Shawn Martin, a Penn Scholarly Communication Librarian, explained, “Penn pays at least twice for one article, but can pay up to 7 or more times for the same content,” in the process of hiring researchers to create the content, buying subscriptions from journals, and paying for reuse rights. To be fair, the transition phase from print to digital media has been costly for publishers because they have had to invest in infrastructure for digital availability while still producing print journals. Many publishers argue that while journal prices may have increased, the price per reader has actually decreased due to a surge in the ease of viewers accessing articles online.

Regardless of whether increasing journal prices was justified, a new model for academic publishing emerged in the 1990s in opposition: open access (OA). There are two ways of attaining open access: Gold OA, when the publisher makes the article freely accessible, and Green OA, which is self-archiving by the author. A few years ago, Laakso et al. conducted a quantitative analysis of the annual publication volumes of Direct OA journals from 1993 to 2009 and found that the development of open access could be described by three phases: Pioneering (1993-1999), Innovation (2000-2004), and Consolidation (2005-2009).2 During the pioneering years, there was high year-to-year growth of open access articles and journals, but the total numbers were still relatively small. OA publishing bloomed considerably from 2000 to 2009, growing from 19,500 articles and 740 journals to 191,850 articles and 4,769 journals, respectively. During the innovation years, new business models emerged. For example, BioMedCentral, later purchased by Springer in 2008, initiated the author charge. In 2004, some subscription-based journals began using a hybrid model, such as Springer’s Open Choice program, which gave authors the option of paying a fee to make their article openly available. During the consolidation phase, year-to-year growth for articles decreased from previous years but was still high, at about 20%.

The introduction of open access journals has sparked fierce and passionate debates among scientists. Proponents of open access believe scientific research should be available to everyone from anywhere in the world. Currently, subscription fees prevent many people from accessing the information they need. With open access, students and professors in low- and middle-income countries, health care professionals in resource-limited settings, and the general public would gain access to essential resources. For instance, Elizabeth Lowenthal, MD, at the Penn Center for AIDS Research, recently published a paper in PLoS One analyzing variables that influence adherence to retroviral drugs in HIV+ adolescents living in Botswana. Her decision to publish open access was because “the article will be of most direct use to clinicians working in Botswana and I wanted to make sure that it would be easy for them to access it.” Open access also provides re-use rights and may facilitate a more rapid exchange of ideas and increased interactions among scientists to generate new scientific information.

However, there may also be some downsides to increased access. Open access may increase the number of articles that people have to sift through to find important studies.3 Furthermore, people who do not know how to critically read scientific papers may be misled by articles with falsified data or flawed experiments. While these papers often get retracted later on, they may undermine the public’s confidence in scientists and medicine. Wakefield’s (retracted) article linking vaccines to autism, for example, may have contributed to the rise of the anti-vaccine movement in the US.4 Furthermore, many open access journals require authors to pay for their papers to be published to offset the cost of publication, and some people have taken advantage of this new payment system to make a profit through predatory journals (a list of predatory OA journals can be found here: http://scholarlyoa.com/publishers/). It is clear though, that the expansion of open access from 1993 until present time suggests that open access can be a sustainable alternative to the traditional model of subscription-based academic publishing.

In addition to facilitating access to scientific articles, the Internet has also created opportunities to improve the peer review process. Peer review was designed to evaluate the technical merit of a paper and to select papers that make significant contributions to a field. Scientists supporting the traditional model of publishing argue that the peer review process in some open access journals may not be as rigorous, and this may lead to the emergence of a “Wild West” in academic publishing. Last year, reporter John Bohannon from Science magazine sent a flawed paper to 304 open access journals, and of the 255 journals that responded, 157 accepted the paper, suggesting little or no peer review process in these journals.5 However, even high-impact journals publish papers with flawed experiments.6 Michael Eisen, co-founder of PLoS, wrote, “While they pocket our billions, with elegant sleight of hand, they get us to ignore the fact that crappy papers routinely get into high-profile journals simply because they deal with sexy topics…. Every time they publish because it is sexy, and not because it is right, science is distorted. It distorts research. It distorts funding. And it often distorts public policy.”7 Nature, for example, published two articles last year about acid-bath stem cell induction, which were later retracted due to data manipulation. However, according to Randy Sheckman, editor-in-chief of eLife, “these papers will generate thousands of citations for Nature, so they will profit from those papers even if they are retracted.”8

With digital communication, peer review for a manuscript could shift from a rigid gate controlled by 3 or 4 people, who might not even be active scientists, into a more dynamic, transparent, and ongoing process with feedback from thousands of scientists. Various social media platforms with these capabilities already exist, including ResearchGate9 and PubMedCommons.10 Some open access journals are using different strategies to address these issues in peer review. eLIFE, for example, employs a fast, streamlined peer review process to decrease the amount of time from submission to publication while maintaining high-quality science. On the other hand, PLoS One, one of the journals published by the Public Library of Science, judges articles based on technical merit alone, not on the novelty.

We polled a few scientists at Penn who had recently published for their thoughts on open access and peer review. Most people did not experience a difference in the peer review process at an open access journal compared to non-open access. The exception was at eLIFE, where reviewers’ comments were prompt, and the communication between reviewers and editors is “a step in the right direction,” according to Amita Sehgal, PhD. To improve the peer review process, some suggested a blind process to help eliminate potential bias towards well-known labs or against lesser-known labs.

The digital revolution is changing the culture of academic publishing, albeit slowly. In 2009, the NIH updated their Public Access Policy to require that any published research conducted with NIH grants be available on PubMed Central 12 months after publication.11 Just last month, the publisher Macmillan announced that all research papers in Nature and its sister journals will be made free to access online in a read-only format that can be annotated but not copied, printed or downloaded. However, only journal subscribers and some media outlets will be able to share links to the free full-text, read-only versions.12 Critics such as Michael Eisen13 and John Wilbanks14 have labeled this change merely a public relations ploy to appeal to demands without actually increasing access. It will be interesting to see if other publishers follow this trend.

Scientific communication has yet to reap the full benefits in efficiency made possible by the Internet. The current system is still less than ideal at furthering ideas and research with minimal waste of resources. But this generation of young researchers is more optimistic and may revolutionize scientific publishing as we know it. “I think [open access is] the future for all scientific publications,” says Bo Li, a postdoc at Penn. “I hope all research articles will be freely accessible to everyone in the world.”

A companion opinion article by Penn PhD student Brian S. Cole can be found here.

This article appeared in the Penn Science Policy January 2015 newsletter



Annie Chen
Michael Allegrezza














1Ware M, Mabe M. (2012) The stm report. http://www.stm-assoc.org/2012_12_11_STM_Report_2012.pdf
2Laakso M, Welling P, Bukvova, H, et al. (2011) The development of open access journal publishing from 1993 to 2009. PLoS ONE.
3Hannay, T. (2014) Stop the deluge of scientific research. The Guardian: Higher Education Network Blog. http://www.theguardian.com/higher-education-network/blog/2014/aug/05/why-we-should-publish-less-scientific-research.
4Wakefield AJ, Murch SH, Anthony A, Linnell, et al. (1998) Ileal lymphoid nodular hyperplasia, non-specific colitis, and pervasive developmental disorder in children [retracted]. Lancet. 351:637-41.
5Bohannon, J. (2013) Who’s afraid of peer review? Science. 342(6154): 60-65. DOI: 10.1126/science.342.6154.60
6Wolfe-Simon F, Switzer Blum J, Kulp TR, et al. (2011) A bacterium that can grow by using arsenic instead of phosphorus. Science. 332(6034) 1163-6. doi: 10.1126/science.1197258.
7Eisen M. (2013) I confess, I wrote the Arsenic DNA paper to expose flaws in peer-review at subscription based journals. it is NOT junk. http://www.michaeleisen.org/blog/?p=1439.
8(2014) Episode 12. The eLIFE podcast. http://elifesciences.org/podcast/episode12
9ResearchGate. http://www.researchgate.net/
10PubMed Commons. http://www.ncbi.nlm.nih.gov/pubmedcommons/
11NIH Public Access Policy Details. http://publicaccess.nih.gov/policy.htm
12Baynes G, Hulme L, MacDonald S. (2014) Articles on nature.com to be made widely available to read and share to support collaborative research. Nature. http://www.nature.com/press_releases/share-nature-content.html
13Eisen M. (2014) Is Nature’s “free to view” a magnanimous gesture or a cynical ploy?. it is NOT junk. http://www.michaeleisen.org/blog/?p=1668
14Van Noorden R. (2014) Nature promotes read-only sharing by subscribers. Nature. http://www.nature.com/news/nature-promotes-read-only-sharing-by-subscribers-1.16460

The Richest Return of Wisdom

 By Brian S. Cole

    The real lesson I’ve gleaned from my time in pursuit of a PhD in biomedical research hasn’t been the research itself; indeed many of my colleagues and I came into the program already equipped with extensive bench experience, but the real eye-opener has been how science is communicated.  When I was an undergraduate, assiduously repeating PCR after PCR that quietly and dutifully failed to put bands on a gel, I just assumed that experiments always worked in the well-funded, well-respected, well-published labs that wrote the papers we read in school.  As an undergraduate, I had implicit trust in scientific publications; at the end of the PhD, I have implicit skepticism.  It turns out I’m not alone.

    The open access movement has taken a new tone in the past year: increasing recognition of the irreplicability1 and alarming prevalence of scientific misconduct2 in highly-cited journals has led to questioning of the closed review process.  Such a process disallows the public access to reviewers’ comments on the work, as well as the editorial correspondence and decision process.  The reality of the publication industry is selling ads and subscriptions, and it is likely that editors often override scientific input by peer reviewers that throws a sexy new manuscript into question.  The problem is the public doesn’t get access to the review process, and closed peer review is tantamount to no peer review at all as far as accountability is concerned.

    For these reasons, our current scientific publication platform has two large-scale negative consequences: the first economic, and the second epistemic.  First, intellectual property rights for publicly funded research are routinely transferred to nonpublic entities that then use these rights for profit.  Second, there is insufficient interactivity within the scientific community and with the public as a result of the silo effect of proprietary journals.  The open access revolution is gaining momentum on the gravity of these issues, but to date, open access journals and publishers have largely conformed to the existing model of journals and isolated manuscripts, and while open access journals have enabled public access to scientific publications, they fail to provide the direly needed interactivity that the internet enables.

    In the background of the open access revolution in science, a 70 year old idea3 about a new system for disseminating scientific publications was realized two decades ago on a publicly licensed code stack4 that allows not just open review, but distributed and continuous open review with real-time version control and hypertext interlinking: not just citations, links to the actual source.  Imagine being able to publish a paper that anybody can review, suggest edits to, add links to, and discuss publicly, with every step of that ongoing process versioned and stored.  If another researcher repeats your experiment, they can contribute their data.  If you extend or strengthen the message of your paper with a future experiment, that can also be appended.  Such a platform would utterly transform scientific publication from a series of soliloquies into an evolving cloud of interlinked ideas.  We’ve had that technology for an alarmingly long time given its lack of adoption by researchers who continue to grant highly cited journals ownership over the work the public has already paid for.

    I’ve kicked around the idea of a Wikiscience5 publication system for a long time with a lot of scientists, and the concerns that came up were cogent and constructive.  In testament to the tractability of a wiki replacement for our system of scientific publication is Wikipedia, one of the greatest gifts to humankind ever to grace the worldwide web.  The distributed review and discussion system that makes Wikipedia evolve does work, and most of us are old enough to remember a time when nobody thought it would.  But how can we assess impact and retain attribution in a distributed publication and review system such as a wiki?  Metrics such as journal impact factor and article-level metrics wouldn’t directly apply to a community-edited, community-reviewed scientific resource.  Attribution and impact assessment are important challenges to any system that aims to replace our journal and manuscript method for disseminating scientific information.  While a distributed scientific information system would not easily fit into the context of the current metrics for publication impact that are an intimate part of the funding, hiring, and promotion processes in academia, the consideration of such a system presents an opportunity to explore innovative analyses of the relevance and impact of scientific research.  Indeed, rethinking the evaluation of scientists and their work6 is a pressing need even within the context of the current publication system.

    We should be thinking about the benefit of the networked consciousness of online collectivism, not the startling failures of our current publication system to put scientific communication into the hands of the public that enabled it, or even the challenges in preserving integrity and attribution in a commons-based peer production system.7  We are the generation that grew up with Napster and 4chan, the information generation, the click-on-it-and-it’s-mine generation, born into a world of unimaginable technological wealth.  Surely we can do better than paywalls, closed peer review, and for-profit publishers.  We owe it to everybody: as Emerson put it, “He who has put forth his total strength in fit actions, has the richest return of wisdom.” 8


This article accompanies a feature piece about scientific publishing in the digital era and also appeared in the Penn Science Policy Group January 2015 newsletter

Brian S. Cole

1Ioannidis, John P. A. "Why Most Published Research Findings Are False."PLoS Medicine 2.8 (2005): E124.
2Stern, Andrew M., Arturo Casadevall, R. Grant Steen, and Ferric C. Fang. "Research: Financial Costs and Personal Consequences of Research Misconduct Resulting in Retracted Publications." ELife 3 (2014)
3Bush, Vannevar. "As We May Think." The Atlantic. Atlantic Media Company, 01 July 1945.
4"MediaWiki 1.24." - MediaWiki. <http://www.mediawiki.org/wiki/MediaWiki_1.24>.
5"WikiScience." - Meta. <http://meta.wikimedia.org/wiki/WikiScience>.
6"San Francisco Declaration on Research Assessment." American Society for Cell Biology. <http://www.ascb.org/dora/>.
7"3. Peer Production and Sharing." <http://cyber.law.harvard.edu/wealth_of_networks/3._Peer_Production_and_Sharing>.
8Emerson, Ralph W. "The American Scholar." The American Scholar. Web. <http://www.emersoncentral.com/amscholar.htm>.

PSPG Newsletter January 2015


The Penn Science Policy Newsletter is out! Click here or on the image below to view.







Tuesday, December 23, 2014

Penn researchers interview HIV-positive adolescents in Botswana to better understand the factors affecting adherence to antiretroviral treatments

Of the more than three million children infected with HIV, 90% live in Africa. As HIV-positive children become adolescents, it is important that antiretroviral treatments are maintained to protect their own health, as well as to safeguard the adolescents from developing resistant strains of HIV and to prevent infection of other individuals.

HIV-positive adolescents’ adherence to these treatments has been identified as a public health challenge for Botswana. However, the assessment tools testing psychosocial factors that are likely associated with poor adherence have been developed in Western countries and their constructs may not be relevant to African contexts. A new study published in PLOS ONE by Penn researchers Elizabeth Lowenthal and Karen Glanz described the cultural adaptation of these assessment tools for Botswana.

The psychosocial assessments investigate factors that may affect adolescents’ adherence to antiretroviral treatments. As Lowenthal summarized, “one of the key reasons why adolescents with HIV have higher rates of death compared with people with HIV in other age groups is that they have trouble taking their medications regularly.”

Researchers looked at the following factors by testing 7 separate assessment scales developed with Western cohorts for their applicability to Botswanan adolescents.
  • Psychological reactance- an aversion to abide by regulations that impose upon freedom and autonomy
  • Perceived stigma
  • Outcome expectancy- whether treatments were expected to improve health
  • Consideration of future consequences- the extent to which adolescents plan for their futures rather than focusing on immediate gains
  • Socio-emotional support- how adolescents receive the social and emotional support they need

The researchers interviewed 34 HIV-positive Botswanan adolescents in depth, sub grouped by age in order to talk about the factors in ways participants could understand.

The study confirmed the construct validity of some assessment tools, but highlighted four areas that caused tools to not relate to participants:
  • Socio-emotional support for the adolescents mostly came from parents rather than peers.
  • Denial of being HIV infected was more common than expected.
  • Participants were surprisingly ambivalent about taking their medicine.

Some of the tools (psychological reactance, future consequences) required major modifications to obtain construct validity for adolescents with HVI in Botswana.The assessment tools were modified during the course of the study based on participant feedback. Future research will test the association between these modified assessment tools and HIV treatment outcomes in order to provide insight into how to best support HIV infected adolescents.

First author Lowenthal suggested that the study could inform studies of adolescent adherence to other treatments as well, stating that “questions that we are able to answer in our large cohort of HIV-positive adolescents will likely be generalizable to other groups of adolescents with chronic diseases.”

-Barbara McNutt 

Monday, December 15, 2014

Penn researchers identify novel therapeutic target for kidney cancer


Kidney cancer, also known as renal cancer, is one of the ten most common cancers in both men and women. The American Cancer Society’s most recent estimates state that of the predicted 63,920 new cases of kidney cancer this year, roughly 20% of  patients will die from the disease. By far, the most common type of kidney cancer is renal cell carcinoma (RCC). The majority of RCCs are clear cell RCCs (ccRCCs), a subtype characterized by metabolic alterations, specifically increased carbohydrate and fat storage. More than 90% of ccRCCs have been found to have mutations in the von Hippel-Lindau (VHL) tumor suppressor gene; however, kidney specific VHL deletion in mice does not induce tumorigenesis or cause metabolic changes similar to those seen in ccRCC tumors. So what additional factors are needed for ccRCC tumor formation and progression? A recent study by Penn researchers published in the journal Nature identified the rate-limiting gluconeogenesis enzyme fructose-1,6-bisphosphatase (FBP1) as a key regulator of ccRCC progression.

To better understand ccRCC progression, the study’s first author, Bo Li, a post-doctoral researcher in the lab of Dr. Celeste Simon, performed metabolic profiling on human ccRCC tumors while also analyzing ccRCC metabolic gene expression profiles. Compared to the adjacent normal kidney tissue, ccRCC tumors had increased amounts of metabolites involved in sugar metabolism and significantly lower expression of carbohydrate storage genes, including FBP1. Further investigation revealed FBP1 expression was reduced in almost all tumor samples tested (>600) and reduced FBP1 expression strongly correlated with advanced tumor stage and poor patient survival. Thus, understanding the role of FBP1 in ccRCCs could significantly impact the treatment of this disease.

How do reduced levels of FBP1 promote ccRCC tumor progression? The authors found that FBP1 depletion in ccRCC cells stimulates growth and relieves inhibition of sugar breakdown (glycolysis), which provides energy for the growing cancer cells. In addition, VHL mutations associated with ccRCCs prevent the degradation of a transcription factor that responds to decreases in oxygen, known as hypoxia-inducible factor α (HIFα), thus stabilizing it. Stabilized HIFα does not cause FBP1 depletion, but its activity is tightly regulated by FBP1. This study emphasized the importance of the interaction between HIFα and FBP1, particularly when glucose and oxygen levels are low, for the formation and progression of the ccRCC.

Why is this work so important? Little is known about how changes in cell metabolism contribute to the formation and progression of ccRCC tumors. As stated by Li, “elucidating how FBP1 impacts the altered metabolic and genetic programs of ccRCC improves our knowledge of the molecular details accompanying ccRCC progression, and identifies novel therapeutic targets for this common malignancy.” Future work may focus on identifying how FBP1 is suppressed and whether reversing FBP1 suppression could improve patient outcomes. 

-Renske Erion