Raul Rodriguez-Esteban

Retraction rates are on the rise

Scientific enterprise is not just a quest for knowledge and truth; it is also a fairly good reflection of the whole spectrum of human behaviour: from genius, passion and jealousy, to mistakes and misconduct. Although new scientific advances and insights are always exciting, the reaction of many scientists to mistakes and misconduct—and to the accompanying retraction of articles—reflects the collapse of a profound belief in the truth‐seeking nature of the ideal scientist: one who is devoid of ordinary human flaws. Recently, there have been several highly publicized retractions in high‐profile journals, which creates a feeling that the integrity of science is in decline. It also raises the question of whether retraction rates for scientific articles are higher than in the past. Here, we show that the latter is indeed the case. We used the Medline database to …

How many scientific papers should be retracted

Published scholarly articles commonly contain imperfections: punctuation errors, imprecise wording and occasionally more substantial flaws in scientific methodology, such as mistakes in experimental design, execution errors and even misconduct (Martinson et al , 2005). These imperfections are similar to manufacturing defects in man‐made machines: most are not dangerous but a small minority have the potential to cause a disaster (Wohn & Normile, 2006; Stewart & Feder, 1987). Retracting a published scientific article is the academic counterpart of recalling a flawed industrial product (Budd et al , 1998). However, not all articles that should be retracted are retracted. This is because the quality of a scientific article depends, among other things, on the effort and time invested in quality control. Mechanical micro‐fractures in turbojet components are detected more readily than those in sculptures, as airplane parts are typically subjected to much more rigorous testing for mechanical integrity. Similarly, articles published in more prominent scientific journals receive increased attention and a concomitant increase in the level of scrutiny. This therefore raises the question of how many articles would have to be retracted if the highest standards of screening were universally applied to all journals. PubMed provides us with …

Imitating Manual Curation of Text-Mined Facts in Biomedicine

Text-mining algorithms make mistakes in extracting facts from natural-language texts. In biomedical applications, which rely on use of text-mined data, it is critical to assess the quality (the probability that the message is correctly extracted) of individual facts—to resolve data conflicts and inconsistencies. Using a large set of almost 100,000 manually produced evaluations (most facts were independently reviewed more than once, producing independent evaluations), we implemented and tested a collection of algorithms that mimic human evaluation of facts provided by an automated information-extraction system. The performance of our best automated classifiers closely approached that of our human evaluators (ROC score close to 0.95). Our hypothesis is that, were we to use a larger number of human experts to evaluate any given sentence, we could implement an artificial-intelligence curator that would perform the classification job at least as accurately as an average individual human evaluator. We illustrated our analysis by visualizing the predicted accuracy of the text-mined relations involving the term cocaine.

Figure mining for biomedical research

MOTIVATION Figures from biomedical articles contain valuable information difficult to reach without specialized tools. Currently, there is no search engine that can retrieve specific figure types. RESULTS This study describes a retrieval method that takes advantage of principles in image understanding, text mining and optical character recognition (OCR) to retrieve figure types defined conceptually. A search engine was developed to retrieve tables and figure types to aid computational and experimental research. AVAILABILITY http://iossifovlab.cshl.edu/figurome/.

Looking at Cerebellar Malformations through Text-Mined Interactomes of Mice and Humans

We have generated and made publicly available two very large networks of molecular interactions: 49,493 mouse-specific and 52,518 human-specific interactions. These networks were generated through automated analysis of 368,331 full-text research articles and 8,039,972 article abstracts from the PubMed database, using the GeneWays system. Our networks cover a wide spectrum of molecular interactions, such as bind, phosphorylate, glycosylate, and activate; 207 of these interaction types occur more than 1,000 times in our unfiltered, multi-species data set. Because mouse and human genes are linked through an orthological relationship, human and mouse networks are amenable to straightforward, joint computational analysis. Using our newly generated networks and known associations between mouse genes and cerebellar malformation phenotypes, we predicted a number of new associations between genes and five cerebellar phenotypes (small cerebellum, absent cerebellum, cerebellar degeneration, abnormal foliation, and abnormal vermis). Using a battery of statistical tests, we showed that genes that are associated with cerebellar phenotypes tend to form compact network clusters. Further, we observed that cerebellar malformation phenotypes tend to be associated with highly connected genes. This tendency was stronger for developmental phenotypes and weaker for cerebellar degeneration.

Quantifying the complexity of medical research

MOTIVATION A crucial phenomenon of our times is the diminishing marginal returns of investments in pharmaceutical research and development. A potential reason is that research into diseases is becoming increasingly complex, and thus more burdensome, for humans to handle. We sought to investigate whether we could measure research complexity by analyzing the published literature. RESULTS Through the text mining of the publication record of multiple diseases, we have found that the complexity and novelty of disease research has been increasing over the years. Surprisingly, we have also found that research on diseases with higher publication rate does not possess greater complexity or novelty than that on less-studied diseases. We have also shown that the research produced about a disease can be seen as a differentiated area of knowledge within the wider biomedical research. For our analysis, we have conceptualized disease research as a parallel multi-agent search in which each scientific agent (a scientist) follows a search path based on a model of a disease. We have looked at trends in facts published for diseases, measured their diversity and turnover using the entropy measure and found similar patterns across disease areas. CONTACT raul.rodriguez-esteban@roche.com.

Visualizing evolution and impact of biomedical fields

We describe a new tool for visualization of biomedical scientific trends. The method captures variations in scientific impact over time to allow for a comparison of relative significance and evolution of fields similar to a financial market scorecard. The tool is available at SciTrends (http://www.scitrends.net), depicting the evolution of almost 200 thousand biomedical fields in time. With millions of articles on thousands of topics published in biomedicine, we envision that only with such large-scale tools researchers can objectively understand the ever-changing interests in the biomedical sciences and make more informed decisions.

A recipe for high impact

Our analysis highlights common statistical features of high-impact articles; we also show how information flows among various publication types.

Six senses in the literature: The bleak sensory landscape of biomedical texts

“It is beyond our power to fathom,Which way the word we utter resonates,Thus, like a sudden grace that comes upon us,A gift of empathetic understanding emanates.” Fyodor Tyutchev (1803–1873), Russian poet (Translated by Mikhail N. Epstein at Emory University, Atlanta, GA, USA). When we read prose—whether technical or literary—our mind parses sentences to recover their meaning. Yet, the flow of the words themselves can invoke surprising or unexpected sensory responses, even for the writer. Even a very rational and technical text can typically affect the reader on multiple cognitive levels, in addition to its basic task of transmitting the author‐intended meaning. > Prose in particular can modify an unsuspecting readers physiological and emotional states profoundly Prose in particular can modify an unsuspecting readers physiological and emotional states profoundly. The semantic priming test in modern psychology exploits this phenomenon—for example, people start to feel and behave as though they have suddenly grown older after they have read a scrambled sequence of words enriched with ageing‐related connotations (Srull & Wyer, 1979). The priming effect is largely independent of our conscious understanding of a text: autistic children, whose text comprehension is mildly impaired, respond to semantic priming the same way as non‐autistic children (Saldana & Frith, 2006). Furthermore, our emotional response to a sequence of words depends in part on our genetic background. Children of parents with bipolar disorder, for example, have been shown to react much more vividly to words that have undertones of social threat than children in a control group (Gotlib et al , 2005). Semantic priming can significantly affect the model of the outside world reported by our senses; merely naming an odour ‘cheddar cheese” or ‘body odour” can determine our perception of it as being pleasant or nauseating (de Araujo et al , 2005). The selection of words in a …

Oreja: A MATLAB environment for the design of psychoacoustic stimuli

The Oreja software package (available from www.liv.ac.uk/psychology/Downloads/Oreja.htm) was designed to study speech intelligibility. It is a tool that allows manipulation of speech signals to facilitate study of human speech perception. A feature of this package is that it uses a high-level interpreted scripting environment (MATLAB), allowing the user to load, break down into different channels, analyze, and select the parts of the signal(s) of interest (e.g., attenuation of the amplitude of the selected channels, addition of noise, etc.).