Identify unexpected outcomes, provide alternative hypotheses, and optimize future experiments


  • We use machine learning methods to identify classifiers of drug responders and non-responders based on patient-level clinical/biomarker data.
  • We identify patient characteristics in sub-groups of interest, revealing non-obvious patterns identifying novel relationships between clinical outputs, and thereby generate new hypotheses about disease progression.
  • In addition to simulated outcomes of a novel agent, we provide data analysis for determining:
    • Responder/non-responder to standard of care in a disease population
    • Patterns and preferences across patient-entered data (i.e., from patient-reported outcomes, health or fitness apps, and or survey response entries)
    • Comparisons of food, cosmetic, or drugs based on chemical content and/or patient biomarker data
    • Novel areas for drug development by examining genetics, proteomics, or metabolomics data across patients with a specific indication (as compared to healthy profiles)