Nanopore direct RNA dataset expands early Alzheimer’s transcriptomics in mouse models
A new Alzheimer’s disease resource moves beyond conventional short-read sequencing and into the molecule itself. Investigators assembled an Oxford Nanopore direct RNA sequencing dataset from 24 mouse brain samples drawn from two amyloid-associated models and their matched controls, creating a platform for transcript-level and isoform-level analysis in early disease biology.
The dataset spans six biological replicates per group, with three males and three females in each cohort. Four groups anchor the resource: 8-week-old C57BL/6 control mice, Aβ1–42-injected C57BL/6 mice, 4-month-old C57BL/6 controls, and age-matched APP/PS1 transgenic mice. Native RNA was sequenced on the Oxford Nanopore PromethION platform, which preserves features often lost in short-read workflows, including long-range transcript structure, poly(A)-related information, and direct observation of RNA molecules without the intermediate conversion bias of cDNA.
The technical profile is substantial. Each sample produced approximately 6.5 million to 9.5 million reads and 8.0 to 10.8 Gb of sequence data. Mean read lengths ranged from about 1.2 kb to 1.4 kb, while N50 values ranged from 1.86 kb to 2.23 kb. Read quality remained above Q19, and alignment to the Mus musculus GRCm39 reference genome yielded an average mapping rate of 88.2 ± 3.5%. In practical terms, those metrics place the dataset well within the range needed for downstream transcriptome reconstruction, isoform resolution, and comparative analysis across models.
Two amyloid models, one shared question about early molecular injury
The value of the design lies in the pairing of an acute and a chronic amyloid context. Aβ1–42 injection captures a more immediate amyloid-associated insult, while APP/PS1 transgenic mice represent a sustained, genetically driven disease state. Sequencing both side by side makes it possible to ask whether early transcriptional responses are model-specific or whether they converge on shared molecular programs linked to synaptic dysfunction, inflammation, RNA processing, or neuronal stress.
This matters because Alzheimer’s pathology is not defined only by altered gene abundance. Splicing changes, isoform switching, polyadenylation differences, and RNA modification patterns can reshape protein output without dramatic shifts in total transcript counts. Short-read RNA sequencing often compresses those layers into fragments that are difficult to assign to full-length isoforms. Direct RNA sequencing retains the native transcript architecture and opens a view onto the post-transcriptional machinery that may be active long before overt neurodegeneration becomes visible.
One sentence captures the central gain here: the dataset preserves RNA structure that short-read sequencing routinely blunts.
Reproducibility and group separation support the resource’s analytic value
Principal component analysis and correlation analysis supported strong reproducibility among biological replicates and clear differentiation among experimental groups. For a data descriptor, that is not a minor technical footnote. It establishes that the biological signal is stable enough to support comparative work rather than collapsing under within-group noise.
The reported outputs extend well beyond aligned reads. The resource includes raw signal files, basecalled sequencing data, genome alignments, transcript annotations, transcript abundance matrices, poly(A)-related outputs, and detailed sample metadata. That package is the difference between a static expression snapshot and a flexible discovery platform. Researchers can move from global expression patterns to transcript structure, then to polyadenylation biology, and finally to RNA modification analyses without rebuilding the dataset from scratch.
Because the samples include both sexes and biological replication across all groups, the material is also positioned for analyses that probe variability rather than averaging it away. Sex-linked expression differences, model-dependent signatures, and transcript-level heterogeneity all become accessible when the underlying reads are long enough and the metadata are complete enough to support disciplined stratification.
Why poly(A) tails and isoforms matter in Alzheimer’s biology
Direct RNA sequencing is especially suited to a disease like Alzheimer’s because RNA regulation in the brain is layered. Neurons rely on precise control of transcript localization, alternative splicing, and isoform usage to maintain synaptic function. When amyloid stress disturbs those controls, the transcriptome can change in ways that are invisible to gene-level summaries.
Poly(A)-tail profiling adds another dimension. Tail length and tail-related processing can influence RNA stability, translation efficiency, and transcript fate. In a neurodegenerative setting, those properties may help distinguish transcripts that are being actively stabilized from those headed toward decay. The same logic applies to RNA modification analyses: chemical marks on RNA can alter processing and expression programs, providing another route through which amyloid toxicity may reshape neuronal biology.
The acute Aβ1–42 model and the chronic APP/PS1 model also invite a temporal interpretation. Early injury may trigger stress-response and inflammatory transcripts, while chronic disease may show deeper remodeling of splicing and RNA homeostasis. A direct RNA dataset is built to detect precisely those differences, because it reads complete molecules rather than reconstructing them from overlapping fragments.
A resource built for comparative work, not a single conclusion
This is a data resource rather than a mechanistic claim. Its immediate value lies in what it enables: transcriptome-wide comparisons between acute and chronic amyloid exposure, isoform discovery, alternative splicing analysis, and downstream searches for molecular signatures that might track early Alzheimer’s pathology in mouse brain.
The metadata and sequencing depth make the dataset useful for investigators studying transcript regulation in neurodegeneration, but its interpretive reach still depends on how the community uses it. The article does not claim a disease biomarker, a therapeutic target, or a validated pathway. It provides a structured, reproducible foundation for those questions to be asked with better resolution than short-read approaches usually allow.
The most concrete limitation is also the most important boundary: this is a mouse-brain dataset, and its translational meaning will depend on how closely the observed transcriptomic patterns align with human disease biology. Cross-species comparison, validation in independent cohorts, and functional follow-up will decide how much of the signal reflects early amyloid pathology itself and how much reflects the engineering of the model system. Until then, the dataset stands as a technically rich map of early amyloid-associated RNA biology, not a finished explanation of Alzheimer’s disease.