Swan: a library for the analysis and visualization of long-read transcriptomes

Motivation

Long-read RNA-sequencing technologies such as PacBio and Oxford Nanopore have discovered an explosion of new transcript isoforms that are difficult to visually analyze using currently available tools. We introduce the Swan Python library, which is designed to analyze and visualize transcript models.

Results

Swan finds 4909 differentially expressed transcripts between cell lines HepG2 and HFFc6, including 279 that are differentially expressed even though the parent gene is not. Additionally, Swan discovers 285 reproducible exon skipping and 47 intron retention events not recorded in the GENCODE v29 annotation.

Availability and implementation

The Swan library for Python 3 is available on PyPi at https://pypi.org/project/swan-vis/ and on GitHub at https://github.com/mortazavilab/swan_vis.