J. Voznica

This study compares deep learning architectures for inferring diversification model parameters from phylogenies, finding that CNNs and RNNs with LTTs outperform other methods for CRBD, while BISSE requires topology and tip states.

This paper applies deep learning to diversification analysis using phylogenies, achieving accuracy comparable to likelihood-based methods but with significantly faster computation.

The paper investigates how defective interfering genomes affect wild-type poliovirus replication using a mathematical model and experimental data, finding that DI replication and encapsidation are key factors in determining viral competition outcomes.

The paper identifies two sources of host cell vulnerability to Salmonella infection using image analysis and modeling, linking morphological and molecular parameters to infection success.