Anna Fowler

The paper evaluates kmer-based representations of TCR repertoires for classification, showing that reduced amino acid alphabets improve performance and interpretability in autoimmune disease classification.

This study uses machine learning to predict coronavirus animal hosts from spike protein and whole genome sequence composition biases, achieving 73% accuracy in cross-validation.

This paper introduces a Bayesian model to identify vaccine-specific BCR sequences by analyzing B cell receptor repertoires across time and individuals, achieving 69% sensitivity.