It’s building mathematical models of the physiological and clinical impact of disease and treatment on virtual patients.
Nova’s knowledge-based models comprise:
A disease model, capturing the most relevant physiological pathways and mechanisms behind a particular condition, plus biological markers linked to clinical outcome(s) of interest.
A virtual population of individuals with the condition of interest. This simulated data includes multiple baseline clinical parameters (e.g. age, blood pressure, inflammatory markers) and the variability in these measures seen across real populations.
A treatment model, representing a drug’s molecular and physiological interactions with the body, e.g. gut absorption or blood-brain barrier transfer.
How does nova build its knowledge-based models?
Nova’s modelers identify and capture relevant information from articles and papers within the scientific literature.
This may be text, images or numbers, for instance describing a link between certain molecular entities, or the time course of a receptor-ligand interaction.
These pieces of knowledge are translated into formatted assertions, or claims.
Our biomodeling experts then rate these claims on the strength of evidence supporting them – such as the number of peer-reviewed publications or confirmatory experiments. They can also be traced back to source documentation, enabling active, transparent knowledge curation.
These steps generate a detailed annotated systems-diagram with multiple interlinked pathways.
This structured knowledge model is then translated into mathematical equations and embedded in computer code.
The outputs of these equations are linked in a way that reflects their relationship in human physiology.
This produces the knowledge-based mathematical model of the system of interest. It can be used to predict the course of the disease or the effects of a drug candidate on an individual and/or on an entire virtual population of interest – a simulated clinical trial.