Vincent

To advance our understanding of the genetics of neurological disorders we need to bring genotype and phenotype information together. Information management techniques for genotype data are well known and not particularly difficult to implement. Information management techniques for phenotype data, particularly in neurological disorders, are more difficult. Paper-based systems, such as DIGS, have the limitation that each time we want to look at different aspects of disease presentation, we need to pull the paper questionnaire and examine the answers for the new aspects for each participant of the study and then re-derive our diagnosis. Online systems, such as VISDIGS, address this to an extent, but do not integrate with genotype data.

We have developed the data model of an integrated information system, called Vincent, to bring genotype and phenotype data together for our bipolar study.

The key features of this system are:

• secure web client allowing on-line data collection from our remote clinics
• integration with the VISDIGS systems, allowing full upload of the VISDIGS data for participants into our centralised database
• perform initial diagnosis from VISDIGS data to present preliminary diagnosis
• ability to generate pedigree diagram data for import into diagramming tools such as Cyrillic and GenoPedigree
• integrated marker and genotype data management
• ability to define a cohort to be a subset of all participants
• ability to define the disease model for linkage analysis
• ability to define the diagnostic model to vary the features from VISDIGS that determine affection status
• directly output .pre and .dat files ready for analysis with systems such as Linkage
• integrated sample location management for blood and blood product
• flexible, powerful reporting feature to generate output to suit all research needs
• security model to ensure appropriate access by authorised staff

Participants visit our clinics and are interviewed using DIGS and the data are entered into VISDIGS running on the clinic computer. Using a web interface, the clinic staff export data from VISDIGS and upload it to our central database. Our scientists upload marker and genotype data into the same database. Our scientists can then define a subset of the study participants for analysis, the parameters associated with the disease models (penetrance levels) and how to translate clinical diagnosis to affection status (the disease model). Vincent can then produce .pre and .dat files for the participants in the analysis with all associated details such as allele values for each marker for each participant, overall allele frequencies and penetrance levels. In order to consider different disease presentation, we only need to alter the component that produces the diagnosis, create new .pre and .dat files and rerun the analysis.
This system saves huge amounts of research time and maintains a clean, single-source version of our invaluable research data. It allows us to consider alternate presentation of the diseases we study without the need to hire staff to re-evaluate paper copies of DIGS questionnaires. It integrates our remote clinics. In the future we will look to enable Vincent to produce linkage analysis files and actually run linkage unattended.

Vincent is built with REALbasic and released as a native binary (Win32, MacOS 8/9 and MacOS X) and is also accessible via a web interface. The data is warehoused in Sybase hosted on Solaris technology.