Many clinical research genomics teams have built deep expertise around short-read analysis. They know how to assess data quality, align reads, call variants, weigh evidence, and connect genomic findings to a subject’s phenotype.
Moving to HiFi long-read sequencing to gain a more comprehensive view of the genome builds on that expertise rather than replacing it. Quality control, alignment, variant calling, and interpretation remain the backbone of the workflow, but what changes is how much genomic context each step has to work with. Longer, highly accurate reads can span repeats, structural changes, and nearby variants that short reads may capture only in fragments, leaving less to infer and more evidence to observe directly.
That added context is also why tool choice matters. Some familiar tools and steps carry over directly, but short-read workflows do not always make use of the additional variation and read-level information that HiFi data provide. Tools designed around HiFi reads help clinical research laboratories capture that value.
For a deeper look into how HiFi sequencing characterizes variation that short reads miss, explore our new guide and its published examples, or build hands-on analysis skills with our on-demand HiFi bioinformatics bootcamp.
Explore the guide Register for the bootcamp
Why HiFi sequencing fits complex clinical research questions
Short-read sequencing is a powerful foundation for detecting single nucleotide variants (SNVs) and small insertions and deletions (indels), yet some clinically relevant questions are inherently difficult to answer with short fragments of DNA. A few practical questions can help a laboratory identify where HiFi analysis may add value.
- Can repeat expansions be sized and characterized with current methods?
- Are structural variants or copy number changes hidden in repetitive regions?
- Do genes with pseudogenes or close paralogs create ambiguity in current results?
- Does the interpretation depend on whether candidate variants are in cis or trans?
- Do samples move through multiple sequential assays before a result is reached?
- Could a broader view of a region reveal findings outside the gene originally suspected?
That last question proved especially important for the Molecular Genetics Unit at CHU Nîmes. The team previously investigated repeat expansions a single gene at a time using multiple methods. After fully implementing a targeted long-read workflow in only three months, it expanded its reportable analysis from seven to eleven genes and found that clinically relevant expansions could occur outside the gene initially suspected. Read the CHU Nîmes story.
Familiar analysis steps with purpose-built tools
PacBio offers a range of computational tools for HiFi data, from supported workflows in SMRT Link to individual command-line tools, and the right starting point depends on the assay. Targeted enrichment and PureTarget repeat expansion data each have dedicated analysis workflows in SMRT Link, while human whole-genome sequencing data can run through the PacBio WGS Variant Pipeline, which shows how HiFi long-read analysis can be consolidated into a familiar workflow. Starting from unaligned HiFi BAM files, the pipeline brings together alignment, variant calling, tandem repeat genotyping, phasing, and methylation summarization. The result is a connected set of evidence for interpretation rather than a separate analysis path for every variant class.
Alignment uses pbmm2, a PacBio wrapper around minimap2. Because HiFi reads are long and highly accurate, they can span challenging or repetitive portions of the genome and help place reads where short fragments may map ambiguously. As with any workflow, quality comes first. Laboratories review read and alignment metrics, confirm sample identity, and verify that data meet predefined requirements before moving to variant calling.
From there, the pipeline draws on tools built to use the length and accuracy of each read.
- SNVs and small indels with DeepVariant
- Structural variants and copy number variants with Sawfish
- Tandem repeats with TRGT
- Variants in homologous or paralogous genes with Paraphase
- Mitochondrial variants with Mitorsaw
- Pharmacogenomic diplotypes with pb-StarPhase
- Phased small and structural variants with HiPhase
Many of these tools, including TRGT, Paraphase, and pb-StarPhase, are also available individually for more focused needs. Because command-line tools can be updated frequently, reproducible analysis depends on version control, documented references and parameters, and laboratory-specific validation.
To see the full workflow in action, this video overview of the WGS Variant Pipeline walks through each tool and the outputs it produces, along with how the pipeline can be deployed on local or cloud compute.
Seeing how variants connect
Phasing is one of the clearest examples of how read-level context makes analysis more direct. Consider a subject with two candidate variants in a recessive disease gene. Whether both sit on the same chromosome or on opposite chromosomes can change the interpretation, and with short reads, answering that question may require parental samples, statistical inference, or additional testing. HiFi reads can connect variants across longer genomic distances, and HiPhase uses this information to determine which small and structural variants travel together on the same DNA molecule. Phasing can then become a natural part of variant interpretation rather than an exercise reserved for the most difficult cases.
Not every laboratory needs to report every output on day one. A rare disease research program may start with SNVs, structural variants, and phasing in unresolved cases, while a neurology group prioritizes repeat expansions and a pharmacogenomics team focuses on genes that are hard to resolve with short reads. Each can begin with the variant classes most relevant to its work and expand as the workflow matures.
Making sense of direct methylation data
Methylation is a frequent question for teams new to HiFi analysis for epigenetics. Because HiFi sequencing reads native DNA, 5mC, 5hmC, and 6mA methylation calling is performed on-instrument, and that information is carried with each read without a separate bisulfite sequencing workflow. Within the WGS Variant Pipeline, MethBat, which has replaced pb-CpG-tools, summarizes CpG methylation probabilities so they can be used in downstream analysis alongside variant calls from the same dataset.
The Kivvi tool shows how this layer can come together with other genomic evidence. In facioscapulohumeral muscular dystrophy (FSHD), a complete answer may depend on repeat size, chromosome of origin, allele type, methylation, and variants in related genes. Kivvi uses HiFi whole-genome data to characterize this information together, turning a fragmented testing question into a more complete genomic view.
Interpreting long-read sequencing results with confidence
Generating a variant call is not the same as generating an answer. Bioinformaticians and variant scientists still assess read support, coverage, annotation, population frequency, phenotype fit, inheritance, and any need for orthogonal confirmation. Long-read data give each of these assessments a richer evidence base, allowing analysts to examine reads spanning a structural variant breakpoint, assess repeat length and composition, and evaluate phase for potential compound heterozygous findings.
As always, laboratories should define which outputs are intended for use, establish evidence thresholds, and validate each reportable claim within their own quality system.
How and where to run a HiFi clinical research analysis workflow
Once a workflow is chosen, the next question is where the analysis will run and how results become an interpretable report. Most laboratories take one of two routes, often mirroring how they handle short-read data today.
Teams with in-house bioinformatics expertise can run PacBio command-line tools and workflows, including the WGS Variant Pipeline, with on-premises or cloud compute. Others work with a tertiary analysis provider that accepts PacBio data and supports the secondary and tertiary workflow, from variant calling through interpretation and reporting. Providers such as EpiSign, Geneyx, Golden Helix, and GenomOncology offer options that can fit alongside a laboratory’s existing setup.
For a closer look at how this comes together in practice, a growing number of published studies walk through the tools and workflow steps behind their results, many of them built on the PacBio WGS Variant Pipeline.
- Stanford’s analytical validation of germline small variant detection using the WGS Variant Pipeline
- GenDx’s assessment of sensitivity for difficult-to-detect pathogenic variants in real-world laboratory samples
- Radboud University Medical Center’s study of difficult-to-detect, clinically relevant variants
- Children’s Mercy Kansas City’s report on a long-read genome sequencing test
- Ambry Genetics’ preprint on yield in short-read genome negative cases, which pairs the WGS workflow with Geneyx
Lessons for practical implementation
The CHU Nîmes story offers a useful model. The team selected validation samples with known repeat expansions and involved technicians, bioinformaticians, and clinical scientists in building the analysis and interpretation workflow in order to move from installation to routine rare disease use in approximately three months. Their experience highlights three priorities.
- Choose representative validation samples. The validation set should reflect the variant types, genes, allele sizes, and genomic complexity the laboratory intends to assess.
- Develop the interpretation interface early. Results need to be understandable and reviewable by the people making scientific and clinical decisions, not only by the people running the pipeline.
- Bring the full team into workflow design. Laboratory staff, bioinformaticians, and clinical reviewers should agree on quality thresholds, handoffs, review steps, and reporting rules before routine use begins.
What HiFi analysis could mean for clinical research
For many laboratories, a practical path is not an overnight replacement of every short-read assay but a focused first use case, such as unsolved rare disease cases, repeat expansion disorders, structurally complex genes, or a targeted panel that currently requires multiple reflex tests. From there, teams can build confidence in the data, the analysis, and the interpretation process, then expand as their needs evolve.
HiFi long-read analysis keeps the core steps that clinical research teams already know while bringing repeats, structural variants, phase, and methylation into view from a single dataset, with the supporting evidence visible in the reads themselves. As tools, validation studies, and analysis partners continue to mature, that more complete view could help laboratories consolidate sequential testing, revisit unsolved cases with new evidence, and bring some of the most difficult regions of the genome into routine clinical research. For many of the questions that short reads leave open, long-read analysis offers a path toward less inference and more insight.
For teams deciding where long reads could help first, this guide explores the types of variation that short reads often miss, how HiFi sequencing captures them, and the published studies behind that work. For teams ready to put HiFi data into practice, the bioinformatics bootcamp offers a practical starting point for analysis.