scifAI

scifAI performs systematic characterization and functional prediction of therapeutic antibodies from imaging flow cytometry data to analyze immunological synapse morphology and predict T cell cytokine production.

Key Features:

• Imaging Flow Cytometry (IFC) Data Processing: Preprocessing and feature engineering tailored for Imaging Flow Cytometry (IFC) data to extract single-cell morphological features.
• Explainable Machine Learning: Implements explainable machine learning techniques to interpret model outputs and relate morphological features to functional outcomes such as cytokine production and antibody efficacy.
• Predictive Analysis: Quantitatively predicts T cell cytokine production in vitro across multiple donors and therapeutic antibodies and links predictions to morphological changes under immune stimulation.
• Large-Scale IFC Dataset: Includes a large IFC dataset comprising over 2.8 million images of the human immunological synapse.
• Modular Architecture: Modular computational architecture enabling integration into analysis pipelines and scalable antibody screening and functional characterization.

Scientific Applications:

• Antibody Design and Evaluation: Supports systematic characterization and evaluation of therapeutic antibodies to inform antibody optimization.
• Immunological Research: Enables analysis of class frequency and morphological changes under different immune stimulations to study immune cell interactions at the immunological synapse.
• Personalized Medicine: Supports development of individualized therapeutic strategies by predicting donor-specific T cell cytokine responses to antibodies.

Methodology:

Workflows include IFC data preprocessing, feature engineering, explainable machine learning, and quantitative predictive modeling of T cell cytokine production across donors and therapeutic antibodies with linking of predictions to morphological features.