Integrating Complex Pangenome Graphs

Graph databases are increasingly used to handle complex data pipelines, in which interconnected data is exploited for visualization and analytics. We propose a novel method, PanGraph-DB, for performing complex inter-pangenomic analysis within a graph database. As a case study, we focus on the antibiotic resistance in sequenced genomes. Over the past decade, the volumes of genomic data stored in public databases have grown exponentially, to the point of hindering comparative genomics algorithms. We show that, due to the nature of genomic data, graph databases enable accurate data and metadata analysis, visualization, and comparison across diverse genomes in the pangenomic approach. Families of graph-encoded pangenomes can then be integrated under a common mediated graph schema. The graph data integration allows to visualize and compare several pangenomes, as well as to analyze AntiMicrobial Resistance (AMR) gene niches through a combination of graph queries, whose performance and scalability we study.