Additional file 8: Figure S6. of A high-quality annotated transcriptome of swine peripheral blood

Integration of the de novo and genome-guided assemblies. This diagram shows the steps used to integrate the two assemblies that were described in Figs. 1 and 2, and includes the number of PTs that resulted from each step, where appropriate; the overall goal was to identify those genome-guided assemblies that added information to the de novo assembly. The sense strands of 113,286 spliced PTs from the de novo assembly (Set A) were first determined based on consensus splice site sequences after alignment to the USMARCv1.0 reference genome. The spliced PTs in the filtered de novo transcriptome were then compared to their counterparts in the filtered genome-guided transcriptome (Set B) by using the Bedtools intersect utility and custom Perl scripts. If a spliced PT from the de novo assembly shared at least one intron or exon, or all introns and exons with a spliced PT from the genome-guided assembly mapped on the sense strand, then they were considered overlapping or exactly the same, respectively. In addition, unspliced PTs from the de novo and genome-guided assemblies were directly compared, without considering their sense strands, by using Bedtools intersect utility. PTs from the de novo transcriptome assembly, which did not overlap PTs from the genome-guided transcriptome assembly were claimed to be de novo transcriptome assembly-specific, and vice versa. The final integrated transcriptome consisted of 132,928 PTs, which included all 126,225 PTs from the filtered de novo transcriptome assembly and 6703 PTs specific to the filtered genome-guided transcriptome, including both 5972 spliced and 731 unspliced PTs. (PDF 337 kb)