Global Agricultural Science & Tech. Client
Data Management | Data Science
The client’s business is growing. An existing plant breeding process and genetics-based decision pipeline were no longer able to support the volume and throughput required which could result in products being delayed to market.
Ocelot was brought in to simulate alternate breeding pipelines and to build new decision analytics on top of simulated data sets. Provided evidence that a particular change can make the process faster without sacrificing quality.
- Implemented software models of plant breeding decisions
- Created a model of plant breeding biological processes through a mix of custom and 3rd party software
- Implemented a plant breeding pipeline simulation tool that allows the user to experiment with new plant breeding operational processes and variations on key decisions while measuring impact on costs and product quality
- Implemented a data science model to improve a specific key decision at the start of the breeding pipeline
- Implemented an RShiny application to quickly put the new data science model in the hands of the plant breeders
- All simulation experiment results are written up in R Markdown and versioned in client’s Gitlab
Simulations demonstrated multiple decision and pipeline changes can reduce costs and result in a faster time to market without a reduction in quality of products.
- AWS Parallel Cluster
- Slurm grid engine
- Singularity containers