MAPFX

MAPFX pre-processes raw data from Massively-Parallel Cytometry (MPC) experiments and imputes missing infinity markers using backbone markers as predictors, enabling comprehensive single-cell surface protein analysis.

Key Features:

• Background Correction: Adapts a normal-exponential convolution model to remove noise from electronic baseline restoration and fluorescence compensation on raw intensities.
• Technical Variation Removal: Uses a log-normal model with plate, column, and row factors to remove unwanted technical variation from sources such as well effects at the cell-level while accounting for potential association between biology and unwanted factors.
• Infinity Marker Imputation: Imputes missing infinity markers in wells without those measurements using informative backbone markers as predictors.
• FFC Assay Normalisation: Normalises data from Fluorescence Flow Cytometry (FFC) assays.
• Downstream Analysis Compatibility: Produces completed datasets suitable for clustering and other statistical analyses.

Scientific Applications:

• MPC Experiment Analysis: Processes data from Massively-Parallel Cytometry experiments including BioLegend's LEGENDScreen and BD Lyoplates assays for single-cell quantification of more than 200 surface proteins.
• Infinity Flow Protocol Enhancement: Improves upon the Infinity Flow (Inflow) analysis protocol by performing background correction prior to imputation and removing unwanted variation while preserving biological signals.
• Comprehensive Cell Type Classification: Enables more comprehensive classification of cell types through imputation of unmeasured infinity markers across all cells in all wells distributed across three 96-well plates.

Methodology:

The pipeline performs background correction using a normal-exponential convolution model, removes unwanted technical variation using a log-normal model with plate, column, and row factors, and imputes infinity markers using backbone markers as predictors.