Information Technology Reference
In-Depth Information
primary biological data, have earned bibliomics the dubious honor of “Bad Omics
Word of the Day” [ 11 ]. In the present context, the discussion and use of the term
bibliomics will be specifi c to the process of eliciting knowledge from biomedical
literature using techniques such as those associated with data mining.
In pursuit of identifying new knowledge from the growing corpora of biomedical
literature, bibliomics offers a unique perspective that is essential in the era of big
data. As raw data are produced at increasingly unbelievable rates, the complemen-
tary literature offers a key distillation of at least some of these data. Minimally,
published reports provide a narrative description of the experiments and the original
purpose as well as fi ndings relevant at the time of data generation. New types of
publications are also emerging that even more specifi cally focus on description of
data and associated experimental parameters that put the data into context (e.g . ,
Scientifi c Data [ 12 ]). It is rare that a publication describing an original investigation
does not involve data, therefore literature plays an essential role in mediating the
interpretations about data. Subsequent analyses that may involve the amalgamation
of data sets further extend the meaning that can be conferred from data. Finally,
literature presents a distilled view of data that can itself be mined to identify poten-
tially novel relationships that can either lead to the support, refutation, or generation
of hypotheses. It is this last role that is the primary focus of bibliomics.
5.2
Eliciting Knowledge from Biomedical Literature
Continued advances in the technical and practical ability to generate data at unprec-
edented rates has resulted in an avalanche of data that presents a two-fold challenge:
(1) interpreting the data themselves; and, (2) leveraging interpretations to support
the scientifi c process. In the context of biomedicine, biomedical literature is a pri-
mary source of data interpretations and subsequent application of the interpreta-
tions. It is thus the essential role of literature to serve as the repository of complete
biomedical wisdom. Such a position also implicates the enormity of challenges that
are faced with leveraging biomedical literature to be used subsequently to support
scientifi c endeavors. In part this challenge is due to the historical audience of bio-
medical literature: human readers. As such, the utility of biomedical literature for
“out-of-the-box” knowledge discovery using automated techniques is a Herculean
feat. Whilst there will be increasing sets of biomedical literature that are available
in digital format that are machine usable (e.g . , in a structured format like the eXten-
sible Markup Language [XML]), the most value of biomedical literature is in the
narrative descriptions of how data were used or interpreted in the context of a study
that has encoded nuances that only a human reader is capable of appreciating. This
knowledge paradox, where human readability confers greater conveyance of knowl-
edge than more machine-readable formats, suggests that as there is increased gen-
eration of big data in biomedicine there will be increased need to develop approaches
for leveraging biomedical literature to reveal the potential value of the newly ren-
dered atoms of knowledge.
Search WWH ::




Custom Search