DREME is a motif discovery algorithm designed to identify short, core DNA-binding motifs of eukaryotic transcription factors (TFs) from large ChIP-seq datasets. It is optimized for speed and efficiency, finding multiple, non-redundant motifs with statistical significance. DREME outperforms existing algorithms in discovering cofactor motifs and provides discriminative motif discovery, allowing for the identification of TF binding patterns. For example, in mouse embryonic stem cell (mESC) ChIP-seq data for the TF Esrrb, DREME identified binding motifs for eight cofactor TFs important in pluripotency maintenance, while other algorithms found only two. DREME also demonstrated that Sox2 and Oct4 do not bind as an obligate heterodimer in mES cells. It is faster than many existing algorithms, scales linearly with dataset size, and can analyze very large ChIP-seq datasets in minutes. DREME is available as part of the MEME Suite of motif-based sequence analysis tools. It uses a beam search approach to find regular expression motifs, and it efficiently estimates the significance of motifs using Fisher's Exact Test. DREME was evaluated against several other motif discovery algorithms using ChIP-seq datasets from mESC, mouse erythrocytes, and human lymphoblastoid cell lines. It found significantly more cofactor motifs than other algorithms and was able to identify primary and cofactor motifs in these datasets. DREME also demonstrated the ability to distinguish between different TF binding contexts, such as the preference of Oct4 for different motifs depending on the presence of Sox2. Overall, DREME provides a powerful tool for analyzing ChIP-seq data and discovering TF binding motifs.DREME is a motif discovery algorithm designed to identify short, core DNA-binding motifs of eukaryotic transcription factors (TFs) from large ChIP-seq datasets. It is optimized for speed and efficiency, finding multiple, non-redundant motifs with statistical significance. DREME outperforms existing algorithms in discovering cofactor motifs and provides discriminative motif discovery, allowing for the identification of TF binding patterns. For example, in mouse embryonic stem cell (mESC) ChIP-seq data for the TF Esrrb, DREME identified binding motifs for eight cofactor TFs important in pluripotency maintenance, while other algorithms found only two. DREME also demonstrated that Sox2 and Oct4 do not bind as an obligate heterodimer in mES cells. It is faster than many existing algorithms, scales linearly with dataset size, and can analyze very large ChIP-seq datasets in minutes. DREME is available as part of the MEME Suite of motif-based sequence analysis tools. It uses a beam search approach to find regular expression motifs, and it efficiently estimates the significance of motifs using Fisher's Exact Test. DREME was evaluated against several other motif discovery algorithms using ChIP-seq datasets from mESC, mouse erythrocytes, and human lymphoblastoid cell lines. It found significantly more cofactor motifs than other algorithms and was able to identify primary and cofactor motifs in these datasets. DREME also demonstrated the ability to distinguish between different TF binding contexts, such as the preference of Oct4 for different motifs depending on the presence of Sox2. Overall, DREME provides a powerful tool for analyzing ChIP-seq data and discovering TF binding motifs.