teff

teff estimates individualized treatment effects from high-dimensional transcriptomic (gene expression) data using random causal forests to identify patient profiles most likely to benefit from specific treatments.

Key Features:

• Random Causal Forests: Implements random causal forests for causal inference in high-dimensional transcriptomic datasets to estimate individualized treatment effects.
• Targeted Patient Profiling: Identifies patient subgroups predicted to experience differential treatment benefit based on gene expression profiles.
• Application to Psoriasis: Extracted profiles indicating high predicted benefit from brodalumab and identified associations with higher T cell abundance in non-lesional skin at baseline and reduced response to etanercept, corroborated by independent studies.
• High-Dimensional Data Handling: Addresses challenges of multicollinearity and interaction effects inherent in high-dimensional gene expression data.

Scientific Applications:

• Personalized Medicine: Enables estimation of individualized treatment effects to inform treatment selection based on transcriptomic profiles.
• Immunology and Dermatology (Psoriasis): Supports discovery of molecular and cellular predictors of treatment response such as T cell abundance in skin.
• Targeted Therapy Development: Facilitates identification of patient subgroups for targeted therapeutic strategies and biomarker discovery.

Methodology:

Applies causal inference via random causal forests to high-dimensional gene expression/transcriptomic data and explicitly accounts for multicollinearity and interaction effects in estimating individualized treatment effects.