gespeR

gespeR reconstructs gene-specific phenotypes from siRNA screening data using a regularized linear regression model to mitigate off-target effects and improve gene-level interpretation in RNAi functional genomics.

Key Features:

• Statistical modeling: Implements a regularized linear regression model to estimate the phenotype of each siRNA from both on-targeted and off-targeted gene effects.
• Data scale: Demonstrated on 115,878 siRNAs, including single and pooled reagents, from three companies across three pathogen infection screens.
• Image-based phenotype deconvolution: Deconvolutes image-based phenotypes to separate gene-specific signals from confounding off-target effects.
• Reproducibility enhancement: Improves reproducibility between independent siRNA sets targeting the same genes by accounting for off-target contributions.
• Biological validation: Prioritized genes were validated as components of pathogen entry mechanisms and TGF-β signaling pathways.

Scientific Applications:

• RNAi functional genomics: Mitigating off-target confounding to enhance accurate attribution of phenotypes to specific genes in RNAi screens.
• Pathogen infection studies: Identifying and prioritizing genes involved in pathogen entry mechanisms.
• Signaling pathway analysis: Investigating components of complex signaling pathways such as TGF-β.

Methodology:

Uses a regularized linear regression model to estimate siRNA phenotypes from both on-target and off-target gene effects and deconvolutes image-based phenotypes.