Large-cohort cDNA sequencing advances multiomic insights into neurodegenerative disease

Neurodegenerative conditions such as Alzheimer’s disease and Parkinson’s disease impact millions of people globally. For polygenic conditions like these, featuring high inter-individual genetic diversity, only sequencing at great scale can capture the range of variants that contribute to disease. This makes large-cohort research crucial in uncovering the molecular mechanisms behind such diseases, as well as potential biomarkers. At the National Institutes of Health (NIH) Center for Alzheimer’s and Related Dementias, USA, Kimberley Billingsley and her team are deploying multiomic Oxford Nanopore sequencing at scale to gain new insights into these complex conditions.

To date, NIH researchers have sequenced over 3,000 whole human genomes from brain samples collected from a neurodegenerative disease cohort using high-throughput nanopore sequencing. The unrestricted read lengths, together with the ability to sequence native DNA, allowed the team to uncover novel genomic variants ranging from single nucleotide variants (SNVs) to structural variants (SVs), as well as cell-type-specific methylation data from these clinical research samples^1,2. However, understanding the functional effects of these genomic variants, including their potential roles in disease, requires another layer of data.

At the London Calling 2026 conference, Kimberley shared how she and her team turned to the Oxford Nanopore platform for large-cohort sequencing once again — but this time, of cDNA³. By studying whole transcriptomes with long cDNA reads, they will be able to see how the novel variants identified in their whole-genome sequencing data may affect transcript expression in the human brain.

Kimberley described how the transcriptome of the human brain is highly complex. Unravelling this complexity requires going beyond gene-level expression analysis, to isoform-level characterisation. This made Oxford Nanopore sequencing ideal for their needs: long nanopore cDNA reads can span full-length transcripts, allowing confident isoform assignment⁴. In contrast, she explained, reads generated via short-read RNA-Seq often cannot span full-length transcripts, leaving isoform structure to be inferred and resulting in errors in complex loci.

Kimberley shared an example from a cohort of individuals of African ancestry⁵, where nanopore cDNA sequencing allowed the identification of a Parkinson’s disease risk variant that ‘would have been completely invisible with short-read sequencing’³. Next, the group deployed this technique at scale.

To add matched transcriptomic information to their large genomic dataset, the team built a high-throughput, end-to-end pipeline for large-cohort cDNA sequencing of post-mortem brain samples. Starting with extracted RNA, they used the Oxford Nanopore cDNA-PCR Sequencing Kit and a liquid handling robot to automate preparation of up to 24 cDNA libraries at once. They then sequenced on PromethION Flow Cells, generating high outputs of full-length cDNA reads at the scale they required. Finally, they analysed their data with an in-house bioinformatics pipeline to perform transcript identification and quantification. At the time of presenting, the group had applied their pipeline to nearly 200 control brain samples.

As one of the first to use the latest cDNA-PCR sequencing protocol, which is now widely available, Kimberley described how she was ‘really amazed’ at its performance³. They were able to generate median transcript read lengths of 2,000 bp, as well as ‘super consistent’ coverage across transcripts, avoiding the 3’ drop in coverage seen in short-read RNA-Seq. Kimberley also noted that though post-mortem brain samples can vary in RNA integrity, the nanopore data they produced was in line with the quality they would expect from induced pluripotent stem cell (iPSC) lines.

The team’s high-quality cDNA data allowed them to resolve a mean of over 78,000 isoforms per sample, with an average of two isoforms observed per gene. Of the transcripts they identified, 65.5% were missed by short-read RNA-Seq, while 48.5% were missing from the database GENCODE. The transcripts featured novel splice junctions, intron retention events, and untranslated regions.

From the same data, Kimberley and her team were also able to measure transcript poly(A) tail length — a feature that is not possible with standard short-read RNA-Seq protocols.

Analysing data from both brain and cell line samples, they found that cell line transcripts displayed shorter, less variable poly(A) tail lengths than brain transcripts, while nuclear transcripts featured considerably longer tails than mitochondrial transcripts for both sample types. With poly(A) tail length indicating poly(A) tail stability, which is known to play a role in age-related conditions such as Alzheimer’s disease, the group plan to investigate this data further.

Having developed this population-scale isoform discovery technique, the team are now applying it to large cohorts of research samples. By pairing their whole-genome and whole-transcriptome Oxford Nanopore data, they will be able to see the complete picture, lead by understanding of sequence, regulation, and expression. The multidimensional information produced by this powerful, single-platform approach will ultimately further our understanding of the drivers of neurodegenerative diseases. The multidimensional information produced by this powerful, single-platform approach will ultimately further our understanding of the drivers of neurodegenerative diseases.

Oxford Nanopore Technologies products are not intended for use for health assessment or to diagnose, treat, mitigate, cure, or prevent any disease or condition.