IAnimal: a cross-species omics knowledgebase for animals

Area of Science:

• Bioinformatics
• Genomics
• Systems Biology

Background:

• Exponential growth in multi-omics data presents challenges and opportunities for biological interpretation.
• Integrating diverse omics data is crucial for understanding gene regulation and biological systems.

Purpose of the Study:

• To develop IAnimal, a cross-species, multi-omics knowledgebase, to enhance data utilization and simplify multi-omics integration.
• To mine genetic mechanisms underlying objective traits by leveraging integrated omics information.
• To build a deep learning model for extracting gene and trait entities from scientific literature.

Main Methods:

• Curated and integrated 61,191 omics datasets (genome, transcriptome, epigenome) for 21 species.
• Developed a deep learning model using BioBERT and AutoNER to mine gene and trait entities from 2,794,237 abstracts.
• Provided user-friendly web interfaces, APIs, and functional modules for data querying and visualization.

Main Results:

• IAnimal hosts 846.46 TB of clean data across multiple omics layers for diverse species.
• The deep learning model effectively identifies gene and trait relationships from literature.
• Users can easily query, mine, and visualize omics data to understand gene regulation.

Conclusions:

• IAnimal facilitates the comprehensive analysis of biological systems by integrating massive multi-omics data.
• The knowledgebase and associated deep learning model provide valuable insights into gene regulation and trait inheritance.
• IAnimal empowers researchers to explore complex biological mechanisms through accessible multi-omics data exploration.