The study team applied single-nucleus long-read RNA sequencing alongside single-cell long-read methods to profile transcript isoforms in mouse kidney. Investigators sampled kidneys from three healthy adult male C57BL/6J mice and three male mice after bilateral renal ischemia-reperfusion injury, as the authors reported. The combined effort produced a large sequencing resource — a total of 1.3 billion reads — with median read lengths ranging from 757 to 884 bases across samples. Those read-length metrics enabled more confident isoform identification than conventional short-read approaches, the authors noted, because longer reads span exon-exon junctions and transcript ends that short reads typically miss.
The dataset supplied both cell-level and nucleus-level long-read libraries. The preprint states that single-nucleus libraries carried a distinct read composition compared with single-cell libraries, and the authors used those differences to examine how isoform detection and diversity vary by input material.
Data indicated that single-nucleus libraries contained a smaller proportion of reads assigned to annotated isoform models than single-cell libraries; the authors attributed this to a larger fraction of intronic reads in nucleus-derived material. Short sentence: intronic signal dominated nuclei reads. Yet the investigators observed that single-nucleus libraries detected more distinct isoforms and displayed greater isoform expression diversity overall. The team connected this increased diversity, at least in part, to the near absence of mitochondrial gene-derived transcripts in nuclei libraries — mitochondrial transcripts are abundant in whole-cell preparations and can monopolize sequencing depth, thereby reducing isoform discovery sensitivity for nuclear-encoded genes.
Single-cell and single-nucleus libraries therefore offered complementary information. Single-cell long-read data yielded a higher fraction of reads confidently aligned to known isoforms, improving certain assignments. Single-nucleus long-read data, by contrast, expanded the catalog of detected isoforms and captured alternative transcript diversity across cell types. The preprint reports these trade-offs without claiming universal superiority of one approach; instead, authors framed the two as providing different windows on transcript structure.
The single-nucleus analysis comparing healthy and injured kidneys identified cell-type–specific transcriptomic changes. Proximal tubule cells, a major parenchymal epithelial population, showed downregulation of genes associated with oxidative phosphorylation after ischemia-reperfusion injury. The authors did not restrict their observation to gene-level changes; data indicated corresponding shifts in isoform composition for oxidative phosphorylation–associated genes. Short sentence: isoform remodeling accompanied transcriptional downregulation. That dual-level effect — decreased gene expression plus altered isoform usage — suggests that injury reprograms both transcript abundance and the preferred transcript variants within affected pathways.
The preprint emphasizes that isoform-level analysis revealed changes that conventional short-read, gene-level studies would be unlikely to detect. The authors reported that isoform shifts were measurable in the single-nucleus long-read data and used those shifts to refine the interpretation of injury-associated metabolic reprogramming in proximal tubules.
Beyond pathway-level signals, the study team identified specific genes showing marked isoform usage changes after injury. Fut9 and Prrc2b were singled out in the abstract as genes with prominent shifts in isoform composition following ischemia-reperfusion injury. The authors reported these two genes as examples where isoform-specific regulation was detectable in the nucleus-derived long-read data; the abstract does not provide sequence-level isoform annotations or quantitative effect sizes for those changes. Short sentence: Fut9 and Prrc2b surfaced as clear isoform-shift candidates.
The authors did not claim exhaustive functional characterization of these isoform changes in the abstract; they reported discovery and differential isoform usage as primary outputs of the long-read single-nucleus approach. Any mechanistic links between individual isoform switches and downstream cellular phenotypes would require follow-up experiments, which the source did not detail in the abstract.
The preprint status of this work means it has not completed peer review; the authors declared no competing interests. The abstract reports several technical observations but omits some methodological specifics: it does not list exact isoform-assignment thresholds, the precise pipeline for isoform calling, or quantitative effect sizes for individual isoform changes. Short sentence: those details are located in the full manuscript and supplementary files, according to the source.
Single-nucleus libraries produced more intronic reads and fewer reads assigned to annotated isoforms compared with single-cell libraries, a trade-off the authors acknowledged. That read composition raises technical questions about optimal library preparation and bioinformatic filtering for isoform discovery from nuclear RNA — questions the authors did not fully resolve in the abstract. The study also focused on male C57BL/6J mice and bilateral ischemia-reperfusion injury; the authors did not report data for other injury models, sexes, or strains in the abstract. Finally, while the authors reported detection of injury-associated isoform shifts in proximal tubule cells and in genes such as Fut9 and Prrc2b, the abstract did not provide functional validation of those isoform changes.
The manuscript therefore ends on concrete gaps: computational and experimental follow-up are required to map isoform sequences, quantify effect sizes, and link specific isoform switches to cellular function and recovery after injury. Those next steps are documented in the full preprint rather than in the abstract alone.