HyperTraPS: Inferring Probabilistic Patterns of Trait Acquisition in Evolutionary and Disease Progression Pathways

Area of Science:

• Biomedical Sciences
• Computational Biology
• Evolutionary Biology

Background:

• The rapid growth of biomedical data presents challenges in understanding complex biological dynamics.
• Inferring evolutionary and disease progression pathways from large datasets is difficult.

Purpose of the Study:

• To present a generalizable statistical platform, HyperTraPS, for inferring dynamic trait acquisition and loss pathways.
• To enable the analysis of diverse data types and distinguish competing biological mechanisms.

Main Methods:

• Utilizing hypercubic transition path sampling (HyperTraPS), a Bayesian approach.
• Incorporating prior knowledge, quantifying uncertainty, and enabling predictions.
• Applying the method to cross-sectional, longitudinal, and phylogenetic data.

Main Results:

• HyperTraPS efficiently learns progression pathways, distinguishing multiple competing routes.
• The platform identifies parsimonious mechanisms underlying observed data.
• Demonstrated application in ovarian cancer progression and tuberculosis multidrug resistance evolution.

Conclusions:

• HyperTraPS offers a powerful tool for revealing previously undetected dynamic pathways in biological systems.
• The platform facilitates intuitive visualization and prediction of biological trajectories.
• This method enhances our ability to harness large biomedical datasets for scientific discovery.