Daniele Scarcella

This study reviews and meta-analyzes CAR-T cell therapy trials and preclinical models to assess their predictive value, finding that preclinical models poorly predict outcomes when using their own features.

This paper introduces MapQC, a method to evaluate the quality of mapping query datasets to reference atlases, showing it improves the detection of disease-specific cell states.

This paper introduces TouCAN, a contrastive learning method for clustering TCR sequences and predicting antigen specificity, showing comparable performance to existing methods.