Benjamin Kaminow

The paper evaluates how many crystal structures are needed for accurate pose prediction in drug discovery, finding that five structures per scaffold often suffice.

The paper shows that incorporating physics-based priors improves structure-based binding affinity models, achieving performance comparable to ligand-only models on a drug discovery dataset.

This paper describes the development of KinoML, a framework for structure-enabled machine learning in drug discovery, focusing on kinases, and highlights key lessons in data management and model reproducibility.

This study evaluates the impact of using a Pan-human consensus genome on RNA-seq alignment accuracy, finding significant improvements in mapping error reduction and splice junction quantification.

Genome-aligning STARsolo matches simulated truth (R≈0.99) and CellRanger's counts almost exactly, running up to 4.4 times faster than CellRanger; pseudoaligners call hundreds of false-positive genes per cell.