Fig. 3

Generation and characterization of a long-read RNA-seq derived protein database. a Schematic demonstrating grouping of transcript isoforms to protein isoform database entries. Some distinct transcript isoforms may have identical coding regions, producing the same theoretical protein isoform product. b Schematic of the SQANTI Protein classification to compare long-read RNA-seq-derived protein isoforms to those annotated in the reference proteome. c Bar chart showing the frequency of protein isoform classifications for the protein database (total ~ 45,000 entries). d Number of genes in each category described in Fig. 1b–f, classified by the relationship between reference isoforms and predicted sample protein isoforms (genes in high confidence space). e Comparison of the number of sample versus reference isoforms for Subset and Superset isoform comparison scenarios. pFSM, protein full splice match; pNIC, protein novel in catalog; pNNC, protein novel not in catalog