DESSO-DB: A web database for sequence and shape motif analyses and identification

Area of Science:

• Genomics and Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Cis-regulatory motif identification is crucial for understanding gene regulation.
• Deep learning models show promise in motif prediction.
• Accessible web databases are needed to leverage these models.

Purpose of the Study:

• To present DESSO-DB, a web database for motif analysis.
• To provide easy access to identified motifs and analysis tools.
• To facilitate the use of deep learning in motif discovery.

Main Methods:

• Developed DESSO, an in-house deep learning tool for motif prediction.
• Integrated motif prediction results for 690 ENCODE and 1,677 Cistrome DB human ChIP-seq datasets.
• Deployed the database on Google Cloud Platform for public access.

Main Results:

• DESSO-DB provides motif predictions and visualizations for extensive human ChIP-seq data.
• Includes online motif finding and scanning for new datasets (ChIP-seq/ATAC-seq).
• Offers downloadable motif results for various datasets, including cancer and CLIP-seq data.

Conclusions:

• DESSO-DB offers efficient access to motif analysis tools and data.
• The database supports diverse genomic datasets and deep learning-based predictions.
• Freely available resources promote advancements in gene regulatory mechanism research.