WYMM Tour: Lyon, France

Background. Molecular diagnosis has become highly significant for patient management in oncology. Methods. Here, 30 well characterized clinical germline samples were studied with adaptive sampling to enrich the full sequence of 152 cancer predisposition genes. Sequencing was performed on Oxford Nanopore (ONT) R10.4.1 MinION flowcells with the Q20+ chemistry.

Results. In our cohort, 11 samples had Large-Scale Rearrangements (LSR), which were all detected with ONT sequencing. In addition to perfectly detecting the locus of the LSR, we found a known MLPA amplification of exon 13 in the BRCA1 (NM_7294) gene corresponded to a duplication in tandem of both exons 12 and 13 of the reference NM_7300. Similarly, in another sample with a known total deletion of the BRCA1 gene, ONT sequencing highlighted this complete deletion was the consequence of a large deletion of almost 140,000 bp carrying over 5 different genes.

ONT sequencing was also able to detect all pathogenic nucleotide variants present in 16 samples at low coverage. As we analyzed complete genes and more genes than with short read sequencing, we detected novel unknown variants. We randomly selected 6 new variants with a coverage larger than 10X and an average quality higher than 14, and confirmed all of them by Sanger sequencing, suggesting that variants detected with ONT (coverage>10X and quality score > 14) could be considered as real variants.

Conclusions. We showed that ONT adaptive sampling sequencing is suitable for the analysis of germline alterations, improves characterization of LSR, and detect SNV even at low coverage.

Introduction: Structural variants (SV) are genomic rearrangements affecting DNA segments of 50 bp or more. Long-read sequencing (LRS) technologies enable precise identification of complex regions by reducing biases and allowing characterization of DNA or cDNA organization.

Methods: We studied two samples with a copy number gain detected by CGH-array and classified as a variant of unknown significance. For each sample, we performed long-read nanopore sequencing on DNA and cDNA to characterize the rearrangements and investigate the transcriptional consequences.

Results: The analysis of the two copy number gains in our samples, performed using LRS, revealed complex variants: a tandem triplication including part of the ZMYM2 gene in the first sample, and chromoanasythesis associated with a translocation partially involving the RERE gene in the second. The transcriptional impact of these variants on the affected genes was investigated through long-read sequencing of full-length cDNA. This analysis revealed, in Sample 1, truncated ZMYM2 transcripts representing 50% of the total transcripts, and in Sample 2, two aberrant transcripts also accounting for 50% of the total transcripts.

Conclusion: LRS approaches on DNA and transcripts enabled the characterization of complex variants, providing pathophysiological hypotheses regarding their potential impacts. A potential dominant-negative effect is suggested for Sample 1, while haploinsufficiency is suggested for Sample 2. These results highlight the significant impact of third-generation sequencing technologies in cytogenomics.

A large body of evidence suggests that alternative splicing is a crucial process during physiological neurodevelopment. While many studies have shown the importance of individual splice isoforms in the brain, others take a more global approach and have tried to unravel global patterns in alternative splicing in the brain. However, most of these studies are based on short-read sequencing. For our research, we used Oxford Nanopore Technology to understand splicing patterns of full-length transcripts in the cortex at different stages of neurodevelopment in wild type mice. Furthermore, we compare this physiological development with the cortical development of a mouse model that is deficient for splicing co-factor and nuclear speckle protein SON. This is relevant, given that SON deficiency leads to ZTTK syndrome, a broad-spectrum syndrome with a distinct neurodevelopmental component.

Myelodysplastic syndromes (MDS) are a group of clonal myeloid disorders characterized by peripheral cytopenias and myeloid dysplasia. Approximately 50% of MDS cases present with Copy Number Variations (CNVs) at diagnosis, confirming the clonal nature of the disease. Some CNVs are diagnostic, but most contribute to prognostic scoring, particularly within the Revised International Prognostic Scoring System (IPSS-R).

In MDS, thirteen key CNVs are typically assessed using conventional cytogenetics. However, long-read sequencing with Oxford Nanopore Technologies (ONT) appears to be a promising alternative for detecting these abnormalities. This study aims to compare ONT long-read sequencing with conventional karyotyping to determine whether it performs better, worse, or equivalently.

To this end, we retrospectively sequenced 70 bone marrow samples using the PromethION P24. Each sequencing library contained 10 barcoded and pooled samples, prepared following the Ligation Sequencing gDNA – Native Barcoding Kit 24 V14 (SQK-NBD114.24) protocol with the ONT PromethION Native Barcoding 96 V14 kit. Sequencing was performed using R10 M Version flowcells (FLO-PRO114M) for 72–96 hours, with a washing step when the proportion of active pores reached less than 20% (at approximately 44 hours) to maximize pore availability.

We then compared ONT results with conventional karyotyping. Out of 107 CNVs, ONT detected 88, along with 8 additional CNVs not identified by karyotyping. However, 11 CNVs were missed by Nanopore, likely due to sensitivity limitations—where the proportion of cells carrying the anomaly was too low—or insufficient sequencing depth in certain samples.

By further optimizing sequencing performance on the PromethION, these findings remain promising, highlighting both the strengths and limitations of this technique for CNV detection in MDS.

Background: Approximately 5–10% of breast cancers (BCs) are due to inherited genetic mutations. Among germline alterations, in addition to well-known pathogenic single nucleotide variations and small in/dels, large genomic rearrangements (LGRs) also contribute to the mutational landscape. It is therefore mandatory for clinical practice to develop fast and reliable diagnostic methods capable of resolving LGRs for patient management. Methods: We retrospectively analyzed a cohort of high-risk BC patients and their families (83 subjects). We extracted DNA from whole blood and prepared libraries for whole-genome sequencing (WGS) with the Ligation Sequencing DNA V14 kit following the manufacturer's instructions, except for 5 samples that did not meet the protocol requirements, for which we implemented alternative strategies. We sequenced libraries on PromethION P24 and performed data QC using the Nanopore EPI2ME Human Variation workflow. Results: To ensure confident analysis, we aimed at achieving a sequencing depth of 30X (100 Gb) and an N50 greater than 5 kb. In the high-quality sample group, the success rate for depth and N50 was 96% and 98%, with a mean value of 123.6 Gb and 7.3 kb, respectively. In 17 out of 83 samples we needed to re-prepare or re-sequence the libraries. Thanks to optimization strategies we specifically devised, we achieved good results (93.4 Gb and 3.06 kb on average) also for the 5 suboptimal samples. Preliminary analysis of 22 samples confirmed that we achieved the target depth (37.11X on average). Conclusions: We have shown that ONT Ligation kit is suitable for performing long-read sequencing on whole-blood samples, even if stored frozen for a long time, allowing us to obtain data from 83 samples in less than 2 months. Overall, this approach needs optimization, but it could represent a potential valid and rapid alternative for diagnostic workflow for patients with familial breast cancer to resolve cases not associated with classical germline mutations.

Background: Approximately 5–10% of breast cancers (BCs) are due to inherited genetic mutations. Among germline alterations, in addition to well-known pathogenic single nucleotide variations and small in/dels, large genomic rearrangements (LGRs) also contribute to the mutational landscape. It is therefore mandatory for clinical practice to develop fast and reliable diagnostic methods capable of resolving LGRs for patient management. Methods: We retrospectively analyzed a cohort of high-risk BC patients and their families (83 subjects). We extracted DNA from whole blood and prepared libraries for whole-genome sequencing (WGS) with the Ligation Sequencing DNA V14 kit following the manufacturer's instructions, except for 5 samples that did not meet the protocol requirements, for which we implemented alternative strategies. We sequenced libraries on PromethION P24 and performed data QC using the Nanopore EPI2ME Human Variation workflow. Results: To ensure confident analysis, we aimed at achieving a sequencing depth of 30X (100 Gb) and an N50 greater than 5 kb. In the high-quality sample group, the success rate for depth and N50 was 96% and 98%, with a mean value of 123.6 Gb and 7.3 kb, respectively. In 17 out of 83 samples we needed to re-prepare or re-sequence the libraries. Thanks to optimization strategies we specifically devised, we achieved good results (93.4 Gb and 3.06 kb on average) also for the 5 suboptimal samples. Preliminary analysis of 22 samples confirmed that we achieved the target depth (37.11X on average). Conclusions: We have shown that ONT Ligation kit is suitable for performing long-read sequencing on whole-blood samples, even if stored frozen for a long time, allowing us to obtain data from 83 samples in less than 2 months. Overall, this approach needs optimization, but it could represent a potential valid and rapid alternative for diagnostic workflow for patients with familial breast cancer to resolve cases not associated with classical germline mutations.

This study, which was presented orally, was performed by HCL Hospital in collaboration with Oxford Nanopore. It uses long-read sequencing (LRS) to improve the analysis of pharmacogenetics (PGx), with a particular focus on the highly polymorphic CYP2D6 gene. This gene is involved in the metabolism of 20–25% of drugs used in clinical practice. LRS enables the precise detection of star alleles, structural variants, and haplotype phasing. This overcomes the limitations of short-read sequencing (SRS) in resolving hybrid/duplicated alleles, which are critical for predicting phenotypes. The study demonstrates the superiority of LRS over SRS in analysing UGT1A1 promoter repeats and the 43-bp deletion in SLC6A4. There is a growing emphasis on the role of epigenetic regulation, particularly DNA methylation in the CYP2D6 promoter region, which may silence expression independently of genetic variants. This could potentially explain discrepancies between genotype and drug-metabolising phenotypes. Integrating methylation profiling with genetic data could refine PGx predictions, addressing gaps in current models and advancing personalised therapeutics. This dual approach of combining high-resolution genomics and epigenomics positions LRS as a cornerstone for future PGx analysis, optimising clinical outcomes by bridging genetic complexity and regulatory dynamics.