Poster #55 - Jamie Bregman

Isoform-Resolved Single-Cell Transcriptomics with Microfluidics-Free Barcoding and Long-Read Sequencing

Avi Srivastava, PhD, The Wistar Institute Calen Nichols, BS, The Wistar Institute Rajeev Ramisetti, PhD, The Wistar Institute Jamie Bregman, MS, The Wistar Institute

Alternative splicing creates the functional diversity of the human transcriptome through the production of multiple distinct mRNA transcripts from a single gene. Despite this fact, most single-cell RNA-sequencing studies obscure transcript heterogeneity by relying on short-reads that cannot resolve individual isoforms and report only gene-level counts. While single-cell long-read methods can overcome these limitations by leveraging full transcript capture, their high-cost, low-throughput, and complex protocols hinder their widespread adoption in single-cell studies. To address this, we developed Pipoly-seq, a low-cost, microfluidics-free method that isolates cells into droplets seeded with barcoded beads via particle-templated instant partitioning (PIP-seq). The resulting Pipoly-seq library is split into aliquots and can be sequenced in parallel on Illumina (short-reads) and Oxford Nanopore (long-reads), with cellular barcodes being retained across platforms. The sequencing data is then processed with Bagpiper, the first open-source pipeline to accurately process PIP-seq barcodes. Bagpiper performs the extraction of fixed-spacer barcode sequences, aligns reads to a customizable reference genome, and employs an expectation-maximization quantification algorithm to produce single-cell gene- and isoform-level matrices. Benchmarking with the K562 cell line and human PBMCs shows high-quality single-cell metrics, precise gene abundance estimation, and distinct heterogeneity at the level of individual cells. Furthermore, results show that only 25% of the long-read depth is needed to reproduce short-read gene-expression estimates (ρ > 0.95), demonstrating superior quantification efficiency as a result of full transcript capture. Finally, by leveraging our novel assay and pipeline, we developed a gene-count deconvolution strategy for reference atlases to accurately infer single-cell isoform quantification estimates from short-read gene expression alone. Thus, Pipoly-seq provides a scalable and cost-effective platform for concurrent short- and long-read single-cell transcriptomics, enabling isoform-resolved insights into cellular heterogeneity.