Skip to main content

What is jinkō? Core concepts and typical project workflows

Learn more here on what jinkō is and what are typical steps of an in-silico trial project using the platform.

Jinkō's core concepts​

  • 'White-box' simulations: build in-silico trials with full traceability
    From curated data and scientific knowledge behind models to the observation of clinical outcomes,  every asset is linked, humanly-readable and accessible at any time to project members or invited auditors. This is true even when our AI agent, Kōhai, is involved (for models creation, results exploration, knowledge and data curation...), as MIDD experts remain at the center of the process.

  • Collaboration between experts modelers, scientists and trial manager
    While jinkō offers powerful modeling capabilities, it is not its only intent. Jinkō aims indeed at breaking silos between modelers and R&D team, making models, trials and results visualizations accessible to both modeling experts, other scientists (biologists, biostatisticians...) and trial managers. 

  • A combination of expert features and models library to accelerate your research
    Jinkō is the fruit of years of experience from nova's extensive expert team of modelers and specialized scientific software engineers. As a result, jinkō is not only a state-of- the-art modeling & trial simulation software, it also optionally allows to start your project from one of nova's expert models, in research fields such as NSCLC, NASH, HBV... to accelerate your own research with a strong base. 

  • A flexible application with a premium User Interface, an API and an AI agent
    Jinkō comes with a modern UI that facilitate all interactions with its powerful features, encourage collaboration between team members from different groups and provide in-context access to our AI agent Kōhai for faster modeling and simulation. It also offers a complete API to interact with the platform programmatically and make jinkō part of your own software ecosystem.

  • Out-of-the-box parallelization
    Jinkō is designed to take advantage of modern hardware and cloud computing, and run by default simulations in parallel on multiple cores or on a cluster of machines. This allows to run large simulations faster and to explore more scenarios in less time.

Typical workflows​

Two flows are possibles in jinkō depending on your entry point, which can be starting from creating / uploading a new model, or starting from a model from our library: 

Starting from scratch, a flow would look as such: 

  1. Curating scientific data and knowledge to build a model
  2. Build a model
  3. Setup and run a trial 
  4. Observe your results and iterate

While starting from a model from our library, the flow looks like this: 

  1. Review the model, its documentation and scientific foundations
  2. Adapt it to your context of use
  3. Setup and run a trial
  4. Observe your results and iterate

Follow the links above to get more details on each step, and / or get a video overview below: