GENEVESTIGATOR® is targeted to the simultaneous analysis of thousands of genome-wide expression profiles.
Despite this large amount of data, we believe the screening should be fast and it should not just show
canned results on pre-defined data sets but allow the user to select experiments of interest from the
large pool of curated data sets. This poses significant challenges.
To address them we have built the GENEVESTIGATOR® engine.
- Client-Server Architecture - To achieve optimal
responsiveness of the analysis tool, our architecture carefully distributes tasks
between the GENEVESTIGATOR® engine on the server and the powerful Java client.
- Parallelization - The core of the GENEVESTIGATOR® engine is a unique parallelization
scheme that distributes each query to multiple servers. The parallelized software and in-memory implementation provide sub-second query results
in cases where even reading the input data from the disk would take several minutes. An example is the search for biomarkers
where one can identify the genes that are most specifically up-regulated in a set of chosen conditions while not
responding to thousands of other conditions.
- Scaling - The parallel nature of the GENEVESTIGATOR® engine
enables us to scale with the envisioned increase in data content both from microarrays and
next generation sequencing, as well as with the growing user base. Already today, hundreds
of analysis sessions with many large-scale queries are performed each day.