This corpus of texts in English was elaborated by the LiCoRN research team at the Université de Bretagne-Sud. It is composed of 8,945 scientific texts, totaling approximately 33 million occurrences. The texts are drawn from 136 scientific journals published on-line by the independent BioMed Central Editor. The original texts, and hence the corpus, are freely accessible by the public, in other words they are exempt from copyright. The journals published by BioMed Central are of good quality; peer-reviewing guaranteeing their quality. Only Experimental sciences are represented in this corpus of Biology and Medicine publications, covering 62 themes: bioinformatics, genomics, genetics… Medicine is slightly more represented than Biology, however the sub-themes are not equally represented. The scientific articles and other written communications cover more than 90% of the corpus across all themes.
As for the French corpus, the textual structure was annotated according to TEI (XML-TEI P5) norms, allowing the incorporation of information on the subject and the type of document. The research themes (Biochemistry, Blood disorders, Cancer, Cell biology…) and the genre types (survey, debate, database, report, short paper…) were coded using TEI headings. Supplementary information were added automatically, if they already figured in the XML original document. The most frequent research themes represented in this corpus are: Medical Genomics, Genomics, Bioinformatics, Genetics, and Women’s Health. Medical Genomics leads with 1,466 texts, then, Genomics with 1,213 texts, and Bioinformatics totalizing 1,074 texts.