General organization of the pipeline. The four main functions are marked i–iv. (i) The create.ref function converts annotation written in GTF into a local variable for downstream use. Aligned long-reads are subsequently compared against the reference in the filter function. Transcript variants containing incongruent exons are singled out and unannotated splice sites are cataloged. (iv) In the following “novel splice site detection” function, unannotated splice site coordinates are quantified and set against a given fasta format reference to scan for adjacent canonical dinucleotides (GU/AG). This step also allows for optional incorporation of splice-junction reads from corresponding short-read data to raise the fidelity. Novel exons that are deemed true upon manual inspection are then added to the reference variable for repeated cycling through the filter function. (ii) Transcript variants that are congruent with the reference annotation are passed on to an enumeration operation, in which the relative and absolute abundance is determined. (iii) The concluding adjacency matrix displays large-scale consecutive exon connectivity.
