Transforming innovation into utility with practical applications

Thermo Fisher Scientific, Algorithm Development for the Cloud

Argentys has completed a significant project for Thermo Fisher, the world’s largest life science reagent and instrument vendor. With this project, Argentys provided genomics domain expertise and software consulting services. Argentys performed an extensive analytical analysis of genomics tools to support the selection of core technology by Thermo management.  We additionally provided scientific software programming services to develop and deploy genomics algorithms on Thermo Fisher Cloud.  The cloud-based algorithm package upgraded obsolete technology to revitalize an important product line and supported integration across Thermo products. Argentys staff worked alongside Thermo staff to integrate components of the complete system.  Throughout the project, Argentys utilized our FlexStaffing™ approach to provide genomics subject matter expertise for guidance and to deliver complex interpretations to support decision-making and to provide scientific advice to our dedicated core-team.  Following comprehensive standalone tests, Argentys assisted the Thermo team with deployment, acceptance, and verification testing.

Specific activities included:

  • Genomics algorithm evaluation
    Identified algorithms, reviewed scientific publications, identified appropriate evaluation data sets, developed validation framework, performed benchmark testing, produced analysis reports, supported management meetings.
  • Genomics algorithm development
    Performed system analysis and data modeling to design algorithmic pipeline. Implemented C / C++ code using modern software development practices. Developed novel signal processing algorithms to eliminate instrument and technology specific artifacts. Codified biological data elements in software data structures. Developed multi-threaded job control module to coordinate execution of algorithmic pipelines on Amazon cloud.  Performed benchmark testing using gold-standard data sets for statistical tuning of algorithmic parameters. Supported client integration testing.