Description

This track displays genome-wide R-loop-forming regions predicted by DeepER, a deep learning-enhanced tool built on a residual BiLSTM architecture. For each 5-kb genomic segment, DeepER computes base-level probabilities of R-loop formation, which are then processed using a sliding window (200 bp window size, 10 bp step size). Consecutive windows with an average probability ≥ 0.95 (default cutoff) are merged to define high-confidence R-loop zones. The track provides both a continuous probability signal (available as a bigWig) and discrete R-loop zones, enabling users to assess predicted R-loop occupancy across the genome. This computational track complements experimental mapping data and offers a standardized view of sequence-intrinsic R-loop formation potential.