Pioneering Secure IT and Business Innovation

Solution:

Built a Streamlit App deployed to the AWS cloud. The app utilized an existing PostgreSQL database and included both static and interactive graphics.

Results:

High-quality PowerPoint-ready graphics saved time for the science team. The ability to quickly access current and historical data resulted in novel research directions. Providing support for the ETL team to identify data inconsistencies added by users.

Solution:

Prepared a Bash pipeline for Next Generation Sequencing data analysis on HPC servers. The pipeline utilized raw sequencing data (NGS, WGS), genotyping data (SNP chips of varying resolutions), and phenotypic data from various sources. Existing genomic software and custom Python/R scripts were employed.

Results:

Several regions associated with the number of production-important traits were identified.

Solution:

Prepared a Python-based ETL process to check data quality, clean errors, and generate a nested JSON file ready for API consumption based on data from questionnaires (CSVs).

Results:

The need for manual data checks and cleaning was minimized. Custom conversion to JSON format enhanced existing ETL processes within the team.

Solution:

• Implementation of Neural Networks for automated tissue segmentation and classification.
• Combining Neural Networks results with classical computer vision and statistical algorithms to create a fully automated data analysis pipeline.

Results:

• Reduction of single image analysis time to several seconds.
• Enhanced capability to analyze historical data.
• Elimination of human factors, resulting in more stable and repeatable analyses.