Fig. 1

Overview of resource and grid design. a We use the Monorail system (detailed in panel b) to analyze and summarize data from archived sequencing datasets, yielding the recount3 and Snaptron resources. recount3 consists of five types of data summaries — Quality Control (QC), gene-level quantification, exon-level quantification, junction counts, and per-base coverages — packaged into tables with associated metadata and organized at the study level. The Snaptron resource consists of the same summaries, but organized at the level of a “collection,” where a single collection might include many studies, and indexed in a way that permits fast queries. Bottom: a few common ways that users can query and interact with the resources. QC, gene and exon-level summaries are compiled using gene annotations, but junction and bigWig coverage summaries are compiled without use of an annotation. b Illustration of Monorail’s grid design, split into Orchestration, Analysis and Aggregation layers. Appendix: Figs. 6, 7, 8, 9 and Appendix: Monorail details contain details about Monorail and its analysis and aggregation components