Sensitivity Analysis & Uncertainty Quantification (UQ) — Hands-on

Develop a solid, practice-oriented understanding of sensitivity analysis and uncertainty quantification (UQ) for systems biology models. Starting from local sensitivities and one-at-a-time (OAT) analyses, you will move to global variance-based methods (Morris, Sobol) and full uncertainty propagation with Monte Carlo and Latin hypercube sampling. The focus is on building robust, reproducible pipelines in Python/R and COPASI for ODE, logical and metabolic models, and communicating results via clear plots and reports.

Sensitivity Analysis & Uncertainty Quantification (UQ)

Help Desk · WhatsApp

Session Index

Session 1 — Concepts & Local Sensitivity Analysis Session 2 — Global Sensitivity: Morris, Sobol & Variance-Based SA Session 3 — Uncertainty Quantification & Propagation Session 4 — Mini Capstone: SA & UQ for a Systems Model

Session 1

Fee: Rs 8800

Concepts & Local Sensitivity Analysis

Why sensitivity & UQ for systems models?

model credibility & robustness parameters vs outputs risk & decision support

Local/derivative-based sensitivity

finite differences normalized sensitivity coefficients one-at-a-time (OAT) analysis

Toolchain & simple workflows

Python: NumPy/SciPy, SALib (intro) R: FME, sensitivity package COPASI local sensitivities

Session 2

Fee: Rs 11800

Global Sensitivity: Morris, Sobol & Variance-Based SA

Sampling strategies for global SA

parameter ranges & distributions Monte Carlo sampling Latin hypercube sampling (LHS)

Morris & Sobol methods

Morris screening (μ*, σ) Sobol first-order & total indices variance decomposition & interpretation

Implementation & visualization

Python SALib R: sensitivity, lhs tornado plots & sensitivity bars

Session 3

Fee: Rs 14800

Uncertainty Quantification & Propagation

Sources & types of uncertainty

parameter vs structural uncertainty measurement noise scenario & model-form uncertainty (overview)

UQ workflows & propagation

Monte Carlo propagation LHS-based UQ prediction bands & credible intervals

Tools & reporting

Python (SALib, SciPy, NumPy) R (FME, sensitivity) COPASI UQ (overview)

Session 4

Fee: Rs 18800

Mini Capstone: SA & UQ for a Systems Model

End-to-end case study: apply SA & UQ to a chosen model (ODE/logic/metabolic)

Theory + Practical

Interpretation & decision support

key drivers & influential parameters robust vs fragile outputs communicating risk & uncertainty

Deliverables

PDF/HTML SA & UQ report Python/R notebook + scripts environment.yml/requirements.txt

PDF

Sensitivity Analysis & Uncertainty Quantification (UQ) Training | Local/Global SA & UQ Workflows

Sensitivity Analysis & Uncertainty Quantification (UQ) — Hands-on