Hani Z. Girgis

EnhancerDetector identifies enhancers across species using interpretable deep learning, showing strong performance and generalizability with experimental validation.

Look4LTRs detects LTR-retrotransposons across multiple plant genomes and identifies recently nested elements, outperforming existing tools in speed and accuracy.

EnhancerTracker identifies cell-type-specific enhancers using an ensemble of deep convolutional neural networks trained on FANTOM5 data, achieving 64% accuracy with three-sequence input.

MeShClust v3.0 improves sequence clustering quality using the mean shift algorithm and alignment-free identity scores, outperforming existing tools on synthetic and real datasets.

FASTCAR predicts pairwise DNA sequence identity scores using alignment-free methods and linear models, achieving faster performance than BLAST and USEARCH on large datasets.

This paper benchmarks alignment-free sequence comparison methods across five research applications and provides a web resource for evaluating and comparing these methods.

Look4TRs detects microsatellites using self-supervised HMMs, outperforming existing tools across eight genomes with higher accuracy and fewer false positives.

LtrDetector detects long terminal repeat retrotransposons de-novo in plant genomes with higher sensitivity and lower false positives than existing tools.

MeShClust2 clusters long bacterial genomes using identity scores without alignment, improving scalability over previous methods.

HebbPlot uses Hebb networks to learn and visualize chromatin mark signatures across 111 epigenomes, revealing distinct patterns for promoters, enhancers, and active genes.

MeShClust uses mean shift and supervised learning to cluster DNA sequences accurately even with imprecise similarity parameters.