Automating the biological data collection process with agents

Scientists spend significant amount of time accessing Web resources, extracting information of interest, filtering, and integrating relevant data from multiple heterogeneous Web sites to support their data collection needs. This tedious collection process is typically performed manually as available technology does not allow scientists to explore and control their data collection process step by step. However, most of the process can be automated. While scripts (e.g., Perl) may be written to retrieve, parse and extract data of interest, many scientists are not programmers and do not have IT support. In contrast we propose an approach based on personal information agents (PIA) that provide scientists a user-friendly mechanism to automate their data collection processes without the need of any programming. This approach is currently being evaluated at the Brain Tumor Cancer Unit of the Translational Genomics Research Institute (TGen), Phoenix, Arizona.