Our partner, a precision diagnostics company, needed to verify and productionize an ML model, developed via an academic collaboration, analyzing proteomic and demographic data to support a precision diagnostic in neurodegenerative disease. They needed a standardized data science workflow to enable the deployment of their assay to a central lab for commercial launch. The model needed to be thoroughly evaluated, through robust testing and sub group analysis, and iteratively refined to be ready for clinical use as an LDT (lab developed test).
Our team developed a comprehensive data science workflow to support the commercial launch of the diagnostic, working with our client’s scientists, their academic partner, and commercial lab partner to ensure our solution met the regulatory and logistic requirements. Quality assurance visualizations were used to monitor the raw data, ensuring data integrity to support validation by the commercial lab. We further optimized the solution for hosting by the commercial lab by containerizing the ML solution, ensuring that it could be deployed seamlessly, providing a robust, reproducible, and user-friendly platform for biomarker data analysis in combination with the proteomic assay to provide high accuracy patient results.