GRAPES-DD: exploiting decision diagrams for index-driven search in biological graph databases

Abstract Background Graphs are mathematical structures widely used for expressing relationships among elements when representing biomedical and biological information. On top of these representations, several analyses performed. A common task is the search one substructure within graph, called target. The problem referred to as one-to-one subgraph search, it known be NP-complete. Heuristics indexing techniques can applied facilitate search. Indexing also exploited in context searching a collection target graphs, one-to-many problem. Filter-and-verification methods that use approaches provide fast pruning graphs or parts them do not contain query. expensive verification phase then performed only on subset promising targets. strategies extract graph features at sufficient granularity level performing powerful filtering step. Features memorized data allowing an efficient access. size, querying time power key points development solutions. Results An existing approach, GRAPES, has been shown have good performance terms speed-up both cases. However, suffers size built index. For this reason, we propose GRAPES-DD, modified version GRAPES which structure replaced with Decision Diagram. Diagrams broad class encode manipulate functions efficiently. Experiments synthetic confirmed our expectation showing GRAPES-DD substantially reduced memory utilization compared without worsening time. Conclusion biochemical completely new potentially thanks their ability compactly sets by exploiting regularity, entire once, instead exploring each single element explicitly. Search based Diagram makes only, more affordable us deal huge ever growing collections structures.