An AI Interface System to NCBI SRA Experiment Data to Support Computational Biology Research

Human language question to relational database interface provides flexible but effective way of querying data because it not only allows people, even with no SQL knowledge to retrieve information from the database, but also makes possible the use of human language in data querying. In this research, we developed an AI interface to SRA database, the meta data of high-throughput next-generation sequencing data made publicly available by NCBI. This system accepts user's data query in natural language, then translates the query into SQL statement using IRNet Deep Learning method which is combined with a Text-based Question Answering method using meta information of large text contents in SRA database. Our system allows computational biologists to query SRA database and retrieving the RNA-seq data for their research without a need for SQL knowledge.