One challenge that researchers face using administrative claims databases from multiple sources is the lack of consistency in the structure of the underlying data.
The STITCH database accumulates data from multiple sources and contains a confidence value for each interaction.
There is a significant risk of having duplicates in the database for cases received from multiple sources.
To construct the protein interaction network, the ANAP database tool utilized data from multiple sources, comprising both predicted interactions and experimentally tested evidence [ 20].
For example, dbNSFP 2.0 is an integrated database of functional annotations from multiple sources for the comprehensive collection of more than 87,361,054 human nonsynonymous SNPs.
We found that PPIs in human, mouse, fly and yeast tend to be recorded in only one database; thus combining data from multiple sources does in fact yield a more comprehensive coverage of known PPIs.
We utilized an ARV market intelligence database comprised of data from multiple sources, including product approvals by the US FDA [ 9, 10] and certifications by the WHO Prequalification Progamme [ 8].
These data are part of the global WHO health workforce database and are derived from multiple sources such as administrative records, establishment census/surveys, labour force or other household surveys, national population, and housing censuses.
This database assembles chemical protein connections from multiple sources, such as ChEMBL (de Matos et al. 2010), BindingDB (Liu et al. 2007), PDSP Ki database (Roth et al. 2004), and PubChem bioassays (Wheeler et al. 2007).
This paper describes the creation of the Canadian Pediatric End-Stage Renal Disease database, a new database formed by combining administrative data from multiple sources to generate a large and rich, longitudinal dataset of Canadian pediatric ESRD patients.
