William Ritchie

The paper develops a method to generate realistic artificial human genomes using adversarial autoencoders, preserving diversity and anonymity while reducing computational demands.

TALC improves long-read transcriptome correction by modeling RNA expression and isoform representation, enhancing downstream RNA-seq accuracy.

S-IRFindeR improves accuracy and consistency in measuring intron retention by addressing biases in existing methods.

The paper shows NF90 modulates processing of a subset of human pri-miRNAs by binding to stable intronic structures, increasing mature miRNA abundance when NF90 is lost.