Clustering more than two million biomedical publications: comparing the accuracies of nine text-based similarity approaches

Background

We investigate the accuracy of different similarity approaches for clustering over two million biomedical documents. Clustering large sets of text documents is important for a variety of information needs and applications such as collection management and navigation, summary and analysis. The few comparisons of clustering results from different similarity approaches have focused on small literature sets and have given conflicting results. Our study was designed to seek a robust answer to the question of which similarity approach would generate the most coherent clusters of a biomedical literature set of over two million documents.

Methodology

We used a corpus of 2.15 million recent (2004-2008) records from MEDLINE, and generated nine different document-document similarity matrices from information extracted from their bibliographic records, including titles, abstracts and subject headings. The nine approaches were comprised of five different analytical techniques with two data sources. The five analytical techniques are cosine similarity using term frequency-inverse document frequency vectors (tf-idf cosine), latent semantic analysis (LSA), topic modeling, and two Poisson-based language models – BM25 and PMRA (PubMed Related Articles). The two data sources were a) MeSH subject headings, and b) words from titles and abstracts. Each similarity matrix was filtered to keep the top-n highest similarities per document and then clustered using a combination of graph layout and average-link clustering. Cluster results from the nine similarity approaches were compared using (1) within-cluster textual coherence based on the Jensen-Shannon divergence, and (2) two concentration measures based on grant-to-article linkages indexed in MEDLINE.

Conclusions

PubMed''s own related article approach (PMRA) generated the most coherent and most concentrated cluster solution of the nine text-based similarity approaches tested, followed closely by the BM25 approach using titles and abstracts. Approaches using only MeSH subject headings were not competitive with those based on titles and abstracts.     

Exosomes communicate protective messages during oxidative stress; possible role of exosomal shuttle RNA

Background

Exosomes are small extracellular nanovesicles of endocytic origin that mediate different signals between cells, by surface interactions and by shuttling functional RNA from one cell to another. Exosomes are released by many cells including mast cells, dendritic cells, macrophages, epithelial cells and tumour cells. Exosomes differ compared to their donor cells, not only in size, but also in their RNA, protein and lipid composition.

Methodology/Principal Findings

In this study, we show that exosomes, released by mouse mast cells exposed to oxidative stress, differ in their mRNA content. Also, we show that these exosomes can influence the response of other cells to oxidative stress by providing recipient cells with a resistance against oxidative stress, observed as an attenuated loss of cell viability. Furthermore, Affymetrix microarray analysis revealed that the exosomal mRNA content not only differs between exosomes and donor cells, but also between exosomes derived from cells grown under different conditions; oxidative stress and normal conditions. Finally, we also show that exposure to UV-light affects the biological functions associated with exosomes released under oxidative stress.

Conclusions/Significance

These results argue that the exosomal shuttle of RNA is involved in cell-to-cell communication, by influencing the response of recipient cells to an external stress stimulus.     

Myocardial Metastases on 6-[18F] fluoro-L-DOPA PET/CT: A Retrospective Analysis of 116 Serotonin Producing Neuroendocrine Tumour Patients

Purpose

This study evaluates the prevalence of cardiac metastases in patients with serotonin producing neuroendocrine tumours (NET), examined with ¹⁸F-FDOPA PET/CT, and the relationship of these metastases to the presence of carcinoid heart disease (CHD) based on echocardiography.

Background

CHD occurs in patients with serotonin producing NET. The diagnostic method of choice remains echocardiography. The precise prevalence of cardiac metastases is unknown given the limitations of standard technologies. Nuclear medicine modalities have the potential to visualize metastases of NET.

Methods

All patients who underwent ¹⁸F-FDOPA PET/CT because of serotonin producing NET between November 2009 and May 2012 were retrospectively analyzed. The presence of cardiac metastasis was defined as myocardial tracer accumulation higher than the surrounding physiological myocardial uptake. Laboratory tests and transthoracic echocardiography (TTE) results were digitally collected.

Results

116 patients (62 male) underwent ¹⁸F-FDOPA PET/CT, mean age was 61±13 years. TTE was performed in 79 patients. Cardiac metastases were present in 15 patients, of which 10 patients also underwent TTE. One patient had both cardiac metastasis (only on ¹⁸F-FDOPA PET/CT) and echocardiographic signs of CHD. There were no differences in echocardiographic parameters for CHD between patients with and without cardiac metastases. TTE in none of the 79 patients showed cardiac metastases.

Conclusion

The prevalence of cardiac metastases detected with ¹⁸F-FDOPA PET/CT in this study is 13%. ¹⁸F-FDOPA PET/CT can visualize cardiac metastases in serotonin producing NET patients. There appears to be no relationship between the presence of cardiac metastases and TTE parameters of CHD.     

Estimated Ultraviolet Radiation Doses in Wetlands in Six National Parks

Ultraviolet-B radiation (UV-B, 280–320-nm wavelengths) doses were estimated for 1024 wetlands in six national parks: Acadia (Acadia), Glacier (Glacier), Great Smoky Mountains (Smoky), Olympic (Olympic), Rocky Mountain (Rocky), and Sequoia/Kings Canyon (Sequoia). Estimates were made using ground-based UV-B data (Brewer spectrophotometers), solar radiation models, GIS tools, field characterization of vegetative features, and quantification of DOC concentration and spectral absorbance. UV-B dose estimates were made for the summer solstice, at a depth of 1 cm in each wetland. The mean dose across all wetlands and parks was 19.3 W-h m⁻² (range of 3.4–32.1 W-h m⁻²). The mean dose was lowest in Acadia (13.7 W-h m⁻²) and highest in Rocky (24.4 W-h m⁻²). Doses were significantly different among all parks. These wetland doses correspond to UV-B flux of 125.0 μW cm⁻² (range 21.4–194.7 μW cm⁻²) based on a day length, averaged among all parks, of 15.5 h. Dissolved organic carbon (DOC), a key determinant of water-column UV-B flux, ranged from 0.6 (analytical detection limit) to 36.7 mg C L⁻¹ over all wetlands and parks, and reduced potential maximal UV-B doses at 1-cm depth by 1%–87 %. DOC concentration, as well as its effect on dose, was lowest in Sequoia and highest in Acadia (DOC was equivalent in Acadia, Glacier, and Rocky). Landscape reduction of potential maximal UV-B doses ranged from zero to 77% and was lowest in Sequoia. These regional differences in UV-B wetland dose illustrate the importance of considering all aspects of exposure in evaluating the potential impact of UV-B on aquatic organisms.     

Reduction of Cr(VI) by a Consortium of Sulfate-Reducing Bacteria (SRB III)

A consortium of bacteria with tolerance to high concentrations of Cr(VI) (up to 2,500 ppm) and other toxic heavy metals has been obtained from metal-refinishing wastewaters in Chengdu, People's Republic of China. This consortium consists of a range of gram-positive and gram-negative rods and has the capacity to reduce Cr(VI) to Cr(III) as amorphous precipitates which are associated with the bacterial surfaces. An endospore-producing, gram-positive rod and a gram-negative rod accumulate the most metallic precipitates, and, over time, 80 to 95% of Cr can be removed from concentrations ranging from 50 to 2,000 ppm (0.96 to 38.45 mM). Kinetic studies revealed a first-order constant for Cr removal of 0.1518 h^-1 for an initial concentration of 1,000 ppm (19.3 mM), and the sorption isothermal data could be interpreted by the Freundlich relationship. The sorption was not entirely due to a passive interaction with reactive sites on the bacterial surfaces since gamma-irradiated, killed cells could not immobilize as much metal. When U or Zn was added with the Cr, it was also removed and could even increase the total amount of Cr immobilized. The consortium was tolerant to small amounts of oxygen in the headspace of tubes, but active growth of the bacteria was a requirement for Cr immobilization through Cr(VI) reduction, resulting in the lowering of E_h. Our data suggest that the reduction was via H₂S. This consortium has been named SRB III, and it may be useful for the bioremediation of fluid metal-refining wastes.     

Negotiating the Therapeutic Gap: Prenatal Diagnostics and Termination of Pregnancy in Sri Lanka

In Sri Lanka, termination of pregnancy, other than in extreme circumstances, is strictly illegal. Among the public and large sections of the medical community there is widespread support for some degree of liberalization of the law, particularly where this relates to serious genetic conditions which can be identified prenatally. Tension emerges out of a publicly maintained conservatism on issues of abortion on the one hand and a growing disconnection from unregulated practices of termination in the private sector on the other. Social science approaches have much to contribute when understanding the ‘therapeutic gap’ that opens up and, in particular, the way that local ideas of fate, destiny and how suffering might be ameliorated become blended with the predictive power of genetic testing.