Information Technology Reference
In-Depth Information
in terms of hypothesis formulation and testing relative to such resources, which
remains extremely basic. As has been illustrated, most if not all hypotheses that are
evaluated in the modern scientifi c setting are generated in a low-throughput manner
based upon the intuition or belief systems of an individual or team of investiga-
tors. Despite historical precedence for such approaches, they are discordant with
the modern, high-throughput data types we regularly encounter, and that are being
generated by EHRs, PHRs, sensor technologies and bio-molecular instrumentation
(to name a few of innumerable examples). In response to this challenge, we can look
to a set of core concepts that underlie alternative and high-throughput approaches
that can lead to in silico hypothesis discovery paradigms. These types of methods
employ domain-specifi c conceptual knowledge collections, such as ontologies or
knowledge that can be extracted from the domain literature using machine learn-
ing or natural language processing methods, in order to reason upon and generate
hypothesis corresponding to a data set or data sets in an extremely high throughput
manner, usually realized via the implementation of knowledge-based and intelligent
software agents. While these types of in silico hypothesis discovery methods remain
very early in their development, they also hold great promise in terms of accelerat-
ing the pace, breadth and depth of scientifi c discovery in the “big data” era, and thus
represent a critical dimension of the vision for Translational Informatics.
Discussion Points
What are the major barriers to the generation and testing of hypotheses in large-
scale and/or heterogeneous data sets?
What differentiates procedural, strategic, and conceptual knowledge? How are
these knowledge types related across a continuum of operationalization?
What role can conceptual knowledge collections play in overcoming the preced-
ing barriers?
As an example of an in silico hypothesis discovery method, what considerations
must be addressed when employing Constructive Induction (CI) relative to con-
cept granularity and/or the evaluation of ensuing hypotheses?
When evaluating the output of knowledge-based intelligent agents used for in
silico hypothesis generation, what is the fundamental difference between the
verifi cation versus validation of such constructs?
References
1. Zerhouni EA. US biomedical research: basic, translational, and clinical sciences. JAMA.
2005;294(11):1352-8. PubMed PMID: 16174693.
2. Sung NS, Crowley Jr WF, Genel M, Salber P, Sandy L, Sherwood LM, et al. Central challenges
facing the national clinical research enterprise. JAMA. 2003;289(10):1278-87. PubMed
PMID: 12633190.
3. Payne PR, Johnson SB, Starren JB, Tilson HH, Dowdy D. Breaking the translational barriers:
the value of integrating biomedical informatics and translational research. J Investig Med.
2005;53(4):192-200. PubMed PMID: 15974245.
Search WWH ::
Custom Search