WTKO-CNN: Deep Learning Reveals Sequence Motifs Distinguishing Wild-Type and Knockout ATAC-seq Peaks
Researchers have developed a deep learning model called WTKO-CNN to distinguish between wild-type and knockout DNA sequences in ATAC-seq data. This model utilizes an attention mechanism to identify critical nucleotide positions influencing classification. By analyzing these influential regions, the team discovered novel sequence motifs associated with transcription factor families that differentiate between the two conditions, offering a new method for understanding transcriptional control. AI
IMPACT Provides a new computational method for discovering functional sequence features in genomic data, aiding biological research.