Long-Read Sequencing for Biomarker Discovery

At a glance:

• Why Long Reads
• Biomarker Classes
• Platform Selection
• Workflow
• Bioinformatics
• Applications
• Candidate Validation
• Sample & Entry
• Why CD Genomics
• Case Study
• FAQs
• Sample Deliverables

Biomarker Classes Resolved by Long-Read Sequencing

Structural-Variant and Complex-Locus Candidates

Candidate biomarkers can be driven by insertions, deletions, inversions, duplications, translocations, repeat expansions, or compound rearrangements that are difficult to reconstruct from short fragments. Long reads can span breakpoints and repetitive flanking sequence, providing direct evidence for the architecture of the event. For projects centered on genome-wide discovery, our Human Whole Genome Sequencing workflow provides long-read coverage for genome-scale variant analysis. When the central question is a complex rearrangement, Human Genome Structural Variation Detection provides a focused route for SV discovery and interpretation.

Phased and Haplotype-Linked Candidates

A candidate may depend on whether variants occur in cis or trans, whether a regulatory change sits on the same molecule as a coding variant, or whether multiple variants define a risk- or response-associated haplotype. Long reads phase variants directly across longer genomic intervals, allowing candidate interpretation at the haplotype level rather than as isolated variant calls. This is particularly useful for highly polymorphic loci, duplicated regions, pharmacogenomic genes, and studies where allele-specific effects are biologically relevant.

Full-Length Isoform and Fusion-Transcript Candidates

Gene-level abundance can hide changes in transcript structure. A biologically important signal may be an isoform switch, novel splice junction, alternative transcription start or end site, intron-retaining transcript, or fusion transcript rather than a change in total gene expression. Our Full-Length Transcript Sequencing (Iso-Seq) service captures complete transcript structures for isoform discovery and comparison. Long-read transcriptomics can therefore turn an ambiguous gene-level association into a defined transcript-level candidate that can be carried into targeted validation.

Single-Cell Genotype and Isoform Candidates

Single-cell expression data are powerful for defining cell states, but short-read libraries can lose the full molecular structure needed to connect cell identity with isoforms, fusion genes, or other transcript-linked genetic features. Our Single-Cell Full-Length Transcriptome Sequencing workflow adds full-length transcript information to cell-resolved studies, supporting candidate discovery in heterogeneous systems where a signal may be restricted to a rare cell population or specific cellular state.

Methylation and Native-Modification Candidates

Long-read single-molecule sequencing can connect sequence context and epigenetic state on the same molecules. For DNA-focused studies, Long-Read Sequencing of DNA Methylation supports methylation-aware analysis alongside long-range sequence information. For RNA-focused projects where native transcript information and RNA modification patterns are central to the hypothesis, Nanopore Direct RNA Sequencing can preserve molecular features that are not retained in conventional cDNA-based workflows.

Integrated Biomarker Discovery Workflow

1. Research question and evidence review — We review existing WGS, RNA-seq, single-cell, array, association, or pilot data and define the unresolved molecular question.
2. Candidate class and sample assessment — We determine whether the project is primarily genomic, transcriptomic, epigenetic, single-cell, or multi-layer and evaluate material quality.
3. Platform and library strategy — PacBio HiFi, ONT genomic DNA, full-length cDNA, direct RNA, or a combined strategy is selected according to the required molecular evidence.
4. Long-read sequencing and assay-specific QC — Libraries are sequenced with QC focused on read quality, molecule length, alignment behavior, coverage distribution, and assay-specific metrics.
5. Feature discovery and molecular resolution — We call and characterize SVs, phased variants, isoforms, fusions, methylation patterns, or other planned features using analysis pipelines matched to the data type.
6. Candidate integration and prioritization — Molecular features are linked to phenotype, condition, cell state, pathway, or prior short-read evidence to produce a ranked shortlist for downstream validation.

The workflow image is intentionally presented as one horizontal decision path. Detailed project-specific steps remain in the accompanying text so the process can be adapted without implying a rigid one-size-fits-all protocol.

Biopharma and Biotech Research Applications

Translational and Disease-Mechanism Research

Long-read sequencing can resolve molecular features that explain why a disease-associated locus or expression signal behaves differently across samples. Examples include a structural rearrangement that changes gene dosage, a phased haplotype linked to expression, a disease-linked transcript isoform, or a methylation state associated with a regulatory region. These features can provide a more precise research hypothesis for downstream experiments.

Drug-Response and Mechanism-of-Action Studies

Treatment-response studies often produce broad expression signatures. Long-read follow-up can ask whether the response is associated with a particular isoform, fusion, allele-specific transcript, structural variant, or epigenetic configuration. Candidate markers can then be prioritized for orthogonal testing in additional samples or experimental models.

Target and Pathway Discovery

A novel transcript isoform, fusion, or complex genomic rearrangement can alter protein domains, regulatory architecture, or pathway membership in ways that are missed by gene-level analysis. Long-read sequencing provides the full molecular sequence needed to predict open reading frames, annotate altered domains, and define candidate mechanisms for functional follow-up.

Pharmacogenomic and Highly Polymorphic Loci

Genes with segmental duplications, pseudogenes, repeats, or extensive haplotype diversity can be difficult to characterize with short reads. Long reads can improve molecular resolution across the locus and support phased candidate analysis. Research conclusions remain project-specific and should be validated with an appropriate independent method before broader use.

Cell-State and Immune Biomarker Research

In heterogeneous tissues, a candidate feature may be present only in a specific malignant, stromal, or immune cell population. Long-read single-cell workflows can connect full-length transcripts and genetic features to cell-state information, supporting research into lineage, clonal structure, immune receptor diversity, and cell-type-specific isoform usage.

Sample and Project Entry Requirements

Biomarker discovery projects can enter at different stages. Final requirements depend on the selected long-read assay, so we confirm acceptance criteria during project design rather than applying one universal input specification.

Independent Published Example: Long-Read Genetic Barcodes Track AML Cell Populations

Background. Penter and colleagues developed a long-read sequencing workflow to recover genetic and transcriptomic features from single-cell cDNA libraries, including somatic mutations, mitochondrial mutations, fusion genes, isoforms, T-cell receptors, and chimeric antigen sequences. The study demonstrates how long reads can add genotype-level information to cell-state profiles when short-read single-cell data alone do not provide sufficient molecular resolution.

Methods. The authors applied targeted long-read sequencing to AML/MDS single-cell cDNA libraries. Across nine patients, they targeted 11 recurrently mutated AML/MDS-associated genes and analyzed 18,097 genotyped profiles. They also used mitochondrial variants as natural genetic barcodes to distinguish and longitudinally track cell populations.

Figure 5 from Penter et al., Nature Communications 2024, CC BY 4.0. Long-read detection of mitochondrial genetic barcodes for donor-recipient and leukemic cell tracking.

Results. In Figure 5, the authors showed that mitochondrial mutations detected from single-cell RNA-seq libraries with long-read sequencing could distinguish donor- and recipient-derived cells and support leukemic tracking. The 10685G>A mitochondrial mutation remained detectable in myeloid progenitor cells during treatment even as their frequency decreased; the authors described it as a potential disease marker for that AML case. In another comparison, 1,855 recipient-derived and 532 donor-derived cells were annotated concordantly using mitochondrial markers and expressed-SNP analysis, with disagreement in only 14 cells. These observations illustrate how a long-read-resolved molecular feature can move from sequence detection to cell-population tracking and candidate-marker interpretation.

Conclusion. This is an independent published example, not a CD Genomics customer project or a performance guarantee. It shows the biomarker-discovery principle behind this solution: long reads can resolve molecular features that connect genotype, transcript structure, clonal identity, and cell state, creating more specific candidates for downstream research validation.

Source: Penter L, Borji M, Nagler A, et al. Integrative genotyping of cancer and immune phenotypes by long-read sequencing. Nature Communications. 2024;15:32. Figure 5, CC BY 4.0.

Sample Deliverables

A typical biomarker-discovery report can include:

1. A candidate evidence matrix ranking features by molecular class, sample recurrence, effect direction, and annotation context.
2. Structural-variant locus views with breakpoint-spanning long reads and phased neighboring variants.
3. Full-length isoform or fusion-transcript structures with condition- or cell-type-specific support.
4. Methylation or modification tracks aligned to the candidate locus when native-molecule data are part of the study.
5. A concise candidate shortlist with evidence summary, interpretation boundaries, and recommended validation route.