ITR: A Novel Graph Database Architecture for Mining Discoveries from the Human Genome

University Of California-Los Angeles, Los Angeles CA

Investigators

Abstract

The goal of the Sutter graph database architecture is to help biologists make sense of the human genome, by enabling them to search genome sequence alignments for patterns of functional and structural relationships between genes. Sequence alignments are the key for discovering meaningful connections between diverse biological data, to make sense of the completed Human Genome. Yet no current database is designed to query detailed sequence alignment relationships as is needed. The Sutter architecture is designed to provide a fast, flexible, and intuitive query system for genomic alignment data, based on storing the entire graph database in a set of indexes, enabling direct lookup for any item to find its relationships. By focusing on indexing, Sutter can move away from the fixed, inflexible schema (table structure) of relational systems, while retaining some of the basis of their speed. A major design goal of Sutter is to implement genomic data objects efficiently, enabling it to store a full genome database in RAM, and achieve dramatically faster query performance. Sutter's first application is to serve the genome research community as an online resource for mining single-nucleotide polymorphisms, their effects on protein function, and mapping disease genes.