Long-read sequencing disentangles isoform complexity at allele-specific loci

In recent years, long-read sequencing technologies have detected transcript isoforms with unprecedented accuracy and resolution. However, it remains unclear whether long-read sequencing can effectively disentangle the isoform landscape of complex allele-specific loci that arise from genetic or epigenetic differences between alleles. Here, we combine the PacBio Iso-Seq workflow with the established phasing approach WhatsHap to assign long reads to the corresponding allele in polymorphic F1 mouse hybrids. Upon comparing the long-read sequencing results with matched short reads, we observed general consistency in the allele-specific information and were able to confirm the imprinting status of known imprinted genes. We then explored the complex imprinted Gnas locus known for allele-specific non-coding and coding isoforms and were able to benchmark historical observations. This approach also allowed us to detect isoforms from both the active and inactive X chromosomes of genes that escape X chromosome inactivation. The described workflow offers a promising framework and demonstrates the power of long-read transcriptomic data to provide mechanistic insight into complex allele-specific loci. Overall design: To capture the allele-specific expression landscape through long-read sequencing, we bred C57BL/6J (BL6) females carrying a heterozygous deletion of the proximal A-repeat (Xist -/+) with a CAST/EiJ (CAST) male. We then collected RNA from a whole brain sample from a resultant F1 female (BL6?Xist × CAST) and performed long-read sequencing. The allele-specific analysis was conducted by counting the reads from each allele and calculating a ratio. This analysis was compared to short-read sequencing allele-specific analysis and does not include replicates.