"What I cannot create, I do not understand."
-Richard Feynman
​
Computational modeling allows us to create a digital twin of any complex system, device, or even patient. These digital twins replicate the structure and function of complex biological and clinical systems, enabling us to simulate, test, and predict outcomes in ways that are impossible or impractical in the real world. From understanding disease mechanisms to forecasting treatment response, modeling gives us a safe, scalable, and insightful tool for discovery and decision-making.
We build two main types of computational models:
-
Predictive models (e.g., neural networks): These data-driven models learn patterns from large datasets, helping us forecast patient outcomes, identify early risk factors, and tailor treatments to individual needs.
-
Mechanistic models (e.g., blood flow in the cardiovascular system, molecular dynamics in a cell): These models are grounded in physical and biological principles, allowing us to simulate system behavior, explore disease progression, and test interventions. Mechanistic models can also be fit to real-world data to estimate parameters that aren’t directly measurable, and we use optimization algorithms to identify system settings that produce the most desirable or efficient outcomes.
At KCMC, our services include: statistics; bioinformatics; machine learning; data analysis; mathematical modeling of biomechanical and molecular systems. (See SERVICES and the portfolio below for more information and examples on previously developed models).
​
By combining predictive and mechanistic approaches, we create powerful tools to improve patient care, accelerate biomedical research, and transform health-informatics applications.
Academic Portfolio of selected projects (industry consulting projects are excluded in compliance with intellectual property agreements, see ABOUT US list of publications for additional projects):
