Pacbio Sequencing Workflow
Pacific Biosciences (PacBio) sequencing generates long reads that reveal structural variants, phased haplotypes, and full length transcripts in a single experiment. This guide explains the end to end PacBio workflow for researchers who need to plan, run, and analyze a HiFi or continuous long read (CLR) project. Whether you are new to long read sequencing or transitioning from short read platforms, you will find practical decision points, step by step implementation, quality checks, and interpretations anchored in published resources NCBI Bookshelf. The goal is to help you avoid common pitfalls and produce reproducible results that stand up to peer review.
Long read sequencing has changed how we resolve repetitive regions and detect large genomic rearrangements. The PacBio system relies on zero mode waveguides that record light pulses from single DNA polymerase molecules. This core technology underpins both the older CLR chemistry and the newer circular consensus sequencing (CCS) that yields highly accurate HiFi reads EMBL-EBI Training. Because your specific biological question dictates which mode to choose, we begin with an overview that frames the entire workflow.
At a Glance
| Aspect | Key Details |
|---|---|
| Core output | Long reads (10,25 kb typical for HiFi, up to >100 kb for CLR) |
| Accuracy | HiFi: >99.9% (Q30+), CLR: ~85,92% raw, corrected by polishing |
| Primary applications | Genome assembly, structural variant detection, haplotype phasing, full length transcript sequencing, metagenomics |
| Library prep times | 4,6 hours for HiFi, similar for CLR with size selection |
| Sequencing run time | 15,30 hours for typical SMRT Cell 8M |
| Data output per SMRT Cell | 5,20 Gb (HiFi) or 10,50 Gb (CLR) depending on insert size and polymerase read length |
| Main decision | HiFi for high accuracy and variant calling, CLR for ultra long reads and complex structural variant discovery |
| Key bioinformatics | PacBio ccs, lima, pbmm2, deepvariant (or guppy), optional polishing with polish or medaka |
Core Concepts and Decision Points
PacBio sequencing begins with a SMRTbell library. The library consists of double stranded DNA flanked by hairpin adapters. During sequencing the polymerase reads the same molecule multiple times in a circular fashion. For HiFi reads, the instrument processes these subreads into a single high quality consensus sequence. For CLR mode it reports only the longest continuous pass.
Decision 1: Choose the chemistry. HiFi (circular consensus) delivers base level accuracy comparable to short reads but with lengths of 10,25 kb. CLR mode sacrifices accuracy for maximum read length, often exceeding 100 kb. If your project requires detection of single nucleotide variants (SNVs) and small indels with confidence, use HiFi. If your goal is to span highly repetitive centromeres or assemble large segmental duplications, CLR may be more cost effective. The trade off is that CLR reads require extensive polishing with short read data or iterative rounds of correction Galaxy Training Network.
Decision 2: Select the target insert size. For HiFi, optimal insert sizes range from 10 to 15 kb. Larger inserts reduce the number of passes the polymerase can complete, lowering the final consensus accuracy. For CLR, bigger inserts (20,50 kb or more) are beneficial to leverage the read length advantage. Size selection using BluePippin or SageELF can narrow the distribution and reduce loading bias.
Decision 3: Determine coverage depth. Whole genome HiFi sequencing typically requires 15,30x coverage for human genomes to enable accurate variant calling and assembly. For targeted regions or smaller genomes, 10,15x may suffice. CLR assemblies need 50,100x raw coverage because individual reads are less accurate. Check the latest recommendations in the PacBio community documentation Bioconductor.
Decision 4: Plan for methylation detection. The PacBio instrument records polymerase kinetics that can be converted into methylation calls. If epigenetic information is part of your study, include a control sample without methylation (e.g., whole genome amplified DNA) and process kinetic data through the pb-CpG-tools pipeline. A recent study used this approach to identify methylation outliers at population scale population scale detection of methylation outliers from long read genome sequencing.
Workflow Steps
Below is a practical sequence from library preparation through data submission. Each step includes a quality check to keep the process on track.
Step 1: DNA Extraction and Quality Assessment
High molecular weight (HMW) DNA is critical. Use a gentle extraction method (e.g., phenol chloroform, magnetic beads, or agarose plug based). The DNA should have an A260/A280 ratio of 1.8 to 2.0 and an A260/A230 ratio above 2.0. Check integrity on a pulsed field gel or TapeStation. DNA fragments below 30 kb can reduce library yield and lead to shorter read lengths. The NCBI SRA repository supplies many example datasets that started with HMW DNA from blood or tissue samples NCBI Sequence Read Archive. For repeat rich regions, such as the STRC gene in hereditary hearing impairment, careful extraction and targeted enrichment can improve coverage analysis of STRC variants uses maker mediated refinement of long read sequencing with MLPA.
Step 2: Library Construction
Shear HMW DNA to the target insert size using a Covaris g-TUBE or similar. For HiFi, shear to 10,15 kb. For CLR, keep intact if possible. Repair ends, A tail, and ligate PacBio overhang adapters. Use the template preparation kit (version depends on instrument). Purify using AMPure PB beads. Perform a nuclease digest to remove incomplete SMRTbell molecules. The final library should be quantified by Qubit and sized by TapeStation. A sharp peak at the expected size with minimal small fragment contamination is ideal.
Step 3: Binding and Sequencing
Anneal the primer to the SMRTbell library. Bind the polymerase using the recommended ratio (typically 4:1 polymerase to SMRTbell). Load onto the Sequel II or Revio instrument. Use the appropriate movie time (15,30 hours for HiFi, up to 20 hours for CLR). Monitor the sequencing metrics in real time: number of productive ZMWs, polymerase read length, and signal to noise. Low productivity may indicate binding or loading issues.
Step 4: Primary Data Processing
Post sequencing, run the PacBio software pipeline. For HiFi, use ccs to process subreads into consensus reads. For CLR, skip this step. Remove adapter sequences and barcodes with lima if multiplexed. After demultiplexing, align reads to a reference genome with pbmm2 (for HiFi) or minimap2 (for CLR). Perform base modification detection if needed. The Galaxy Training Network offers ready to use workflows for these steps Galaxy Training Network.
Step 5: Downstream Analysis
Downstream analysis depends on your question. For variant calling, use DeepVariant or PacBio SMRT Link for SNVs and small indels. For structural variants, use SVDSS, pbsv, or Sniffles. For de novo assembly, use hifiasm (HiFi) or Canu with CLR reads. Haplotype aware error correction methods can further improve assembly quality when reads contain multiple haplotypes haplotype aware long read error correction. For metagenomic profiling with full length 16S rRNA gene amplicons, the Emu classifier combined with GTDB database provides the highest taxonomic resolution PacBio full length 16S rRNA gene sequencing processed with Emu and GTDB provides the highest taxonomic resolution for rumen bacteriome profiling.
Quality Checks and Troubleshooting
Check 1: Post CCS quality. For HiFi reads, the recommended minimum quality score is Q20 (99% accuracy). Filter out reads below this threshold. High quality HiFi runs often yield a median Q30 or higher. If quality is low, consider adjusting the --min-passes parameter (default 3) or increasing the --min-rq threshold.
Check 2: Coverage uniformity. After alignment, inspect coverage across the genome. In targeted regions, low coverage may indicate GC bias or library drop out. Normalize by using PCR cycle auto normalization during library preparation to reduce error rates and sequence distribution biases applying PCR cycle autonormalization to PacBio full length 16S rRNA library preparations.
Check 3: Structural variant validation. Long read calls often include false positives from alignment artifacts. Validate a subset of structural variants by PCR or MLPA as done for STRC variants Integrated analysis of STRC variants. Compare results with an orthogonal short read dataset if available.
Check 4: Assembly contiguity. For de novo assembly, check the N50 and L50 statistics. If the assembly is fragmented, you may need more coverage, better size selection, or a different assembler (e.g., switch from hifiasm to Flye for HiFi).
Common Mistakes
Mistake 1: Starting with degraded DNA. Even slight fragmentation reduces library yield and read length. Always verify HMW DNA integrity before library prep.
Mistake 2: Overloading the SMRT Cell. Too many library molecules cause multiple templates per ZMW, resulting in background noise and lower read quality. Follow the recommended loading concentration (typically 0.5,2 pM).
Mistake 3: Ignoring adapter dimer. A small peak at ~120 bp on the TapeStation indicates adapter dimer. This contaminant reduces effective loading and wastes sequencing capacity. A size selection step before binding can remove it.
Mistake 4: Using default filters without checking. Bioinformatics tools have default thresholds that may not suit your data. For example, pbmm2 default mapping quality filters can discard reads with high indel rates, which are common in repetitive regions. Adjust parameters based on read accuracy.
Limits of Interpretation
Long reads do not guarantee error free assemblies. HiFi reads still struggle with homopolymers longer than 10 bases and with regions of extreme GC content (below 20% or above 80%). CLR assemblies require polishing with Illumina data to reach Q40 quality. Furthermore, methylation calls from PacBio kinetics are only accurate at single base resolution when coverage exceeds 20x, lower coverage leads to noisy detection.
Some biological questions cannot be answered solely with PacBio data. For example, phasing of entire chromosomes may require parental or population scale sequencing. Multi sample assembly pipelines like LORA can help but introduce computational complexity LORA a polymorphic multi sample long read assembly pipeline. Always consider integrating short read data for high confidence SNV calling and for validating complex variants.
Frequently Asked Questions
Q1: How long does a PacBio sequencing run take? A typical run on the Sequel IIe or Revio with HiFi chemistry lasts 15,30 hours. CLR runs are similar or slightly shorter. The instrument time excludes library preparation (4,6 hours) and downstream analysis.
Q2: Can I use PacBio data for RNA sequencing? Yes, PacBio Iso-Seq captures full length transcripts without assembly. The workflow uses polyA selection and reverse transcription followed by SMRTbell library prep. HiFi reads provide accurate isoform detection and quantification.
Q3: What is the minimum amount of DNA required? PacBio recommends 1,5 micrograms of HMW DNA for library preparation. Lower amounts can work but may reduce library complexity and increase PCR duplication.
Q4: How do I share my PacBio data publicly? Deposit raw subreads or HiFi reads in the NCBI Sequence Read Archive (SRA) under a BioProject accession. Include sample metadata, library construction details, and basecaller versions for reproducibility.
References and Further Reading
- NCBI Bookshelf: long read sequencing technology overview NCBI Bookshelf
- EMBL EBI Training: practical guide to PacBio data analysis EMBL-EBI Training
- Galaxy Training Network: end to end PacBio workflows Galaxy Training Network
- Bioconductor: R packages for analyzing long read data Bioconductor
- NCBI Sequence Read Archive: repository for PacBio datasets NCBI Sequence Read Archive
- Haplotype aware long read error correction (Algorithms Mol Biol, 2025) PubMed
- LORA: polymorphic multi sample long read assembly pipeline (NAR Genom Bioinform, 2025) PubMed
- Population scale detection of methylation outliers from long read genome sequencing (medRxiv, 2025) PubMed
- Applying PCR cycle autonormalization to PacBio full length 16S rRNA library preparations (mSphere, 2025) PubMed
- PacBio full length 16S rRNA gene sequencing processed with Emu and GTDB provides the highest taxonomic resolution (ISME Commun, 2025) PubMed