# How to Choose Between Trio-Based and Single-Individual Phasing for Your Long-Read Project: A Decision Guide ## Direct Answer and Scope The decision between trio-based phasing and single-individual phasing for a long-read sequencing project comes down to whether you need parent-of-origin information, how accurate your haplotype blocks must be, and what your budget allows for additional sample preparation and sequencing. Trio-based phasing requires sequencing both parents alongside the proband, which roughly triples the sequencing cost for those samples but provides direct experimental evidence for which variants sit on which chromosome copy. Single-individual phasing uses computational methods on one genome, costs less upfront, and works when parental samples are unavailable, but it produces shorter haplotype blocks and cannot assign variants to maternal or paternal origin without additional inference. This guide gives you a decision framework based on research goals, budget, and data availability, with concrete criteria for when each approach is defensible and when you should escalate to a more expensive design. The practical outcome of this article is a structured assessment you can apply before committing to a sequencing strategy. You will learn what records to keep, what quality metrics to monitor, what failure patterns to expect, and when to consult a bioinformatics specialist or genetic counselor. The guidance applies to researchers planning long-read projects on platforms such as Oxford Nanopore or Pacific Biosciences, and it draws on published evidence about phasing accuracy, parent-of-origin analysis, and clinical applications of long-read genome sequencing. ## Understanding Haplotype Phasing in Long-Read Projects Haplotype phasing is the process of determining which genetic variants are physically located on the same chromosome copy. Each diploid individual carries two copies of each autosome, one inherited from each parent, and variants on the same copy are said to be in the same haplotype. Phasing assigns each variant to one of the two chromosome copies, producing a pair of haplotypes for each chromosome. This information matters for many analyses, including identifying compound heterozygous variants, studying allele-specific expression, and determining parent-of-origin effects where the phenotypic impact of a variant depends on whether it came from the mother or the father. Long-read sequencing produces reads that are thousands to tens of thousands of base pairs long, which means a single read can span multiple heterozygous variants. When a read contains two or more heterozygous variants, those variants are physically linked on the same chromosome copy, and that read provides direct phasing evidence. This is the key advantage of long-read platforms over short-read sequencing, where reads are too short to span multiple variants reliably and phasing must rely on statistical inference from population data or on family relationships. The choice between trio-based and single-individual phasing is not a choice between long-read and short-read sequencing. Both approaches can be applied to long-read data. The difference is in how you obtain the phasing evidence. Trio-based phasing uses the genotypes of both biological parents to determine which alleles the proband inherited from each parent. Single-individual phasing uses the long-read data alone, often combined with a reference panel of haplotypes from a population database, to reconstruct haplotypes computationally. The NCBI maintains databases of genetic variation and reference sequences that are commonly used in phasing workflows, and researchers should be familiar with these resources when planning data analysis [1]. The European Bioinformatics Institute offers training materials on data resources and analysis methods that can help laboratory professionals build the skills needed for phasing projects [2]. These official resources provide the foundational context for understanding what data formats and quality metrics matter in practice. ## At a Glance: Trio-Based Versus Single-Individual Phasing The table below summarizes the key differences between the two approaches across the dimensions that matter most for project planning. Use this table as a starting point for your decision, then read the detailed sections that follow for the reasoning and caveats. | Decision Factor | Trio-Based Phasing | Single-Individual Phasing | | --- | --- | --- | | Parental samples required | Yes, both biological parents | No | | Sequencing cost per proband | Approximately three times the single-genome cost | Single-genome cost | | Parent-of-origin assignment | Direct from Mendelian inheritance | Not possible without additional relatives or inference methods | | Haplotype block length | Whole-chromosome phase in most regions | Variable, often shorter blocks depending on read length and variant density | | Best suited for | Clinical diagnosis, parent-of-origin studies, compound heterozygote detection | Population studies, species with no parental samples, exploratory projects | | Main limitation | Requires parental samples, higher cost, longer turnaround | Cannot resolve parent-of-origin, may have switch errors in low-complexity regions | | Data analysis complexity | Requires joint calling across three samples | Simpler per-sample workflow | The accuracy difference between the two approaches is substantial in specific contexts. Statistical phasing methods applied to large population cohorts have achieved very low switch error rates, which measure how often the phasing algorithm incorrectly swaps which haplotype a variant belongs to. In one benchmark study of more than eight million research-consented customers and the UK Biobank, two state-of-the-art phasing methods achieved a median switch error rate of 0.00 percent in trio children and 0.026 percent in British samples, with errors concentrated in regions lacking identity-by-descent coverage [7]. This evidence shows that computational phasing can be highly accurate when large reference panels are available, but it also shows that accuracy depends on population representation and on having relatives in the dataset to provide identity-by-descent information. For long-read projects specifically, the relevant comparison is not between statistical phasing on large cohorts and trio phasing on a single family. The relevant comparison is between using long reads from one individual to phase that individual versus using long reads from three family members to phase the proband. The population phasing evidence is useful because it establishes what accuracy is achievable with computational methods, and it highlights the conditions under which those methods fail, namely regions without identity-by-descent coverage [7]. ## Core Principles of Phasing Accuracy and Cost ### What Determines Phasing Accuracy Phasing accuracy depends on three main factors: the density of heterozygous variants, the length of the sequencing reads or the distance between informative markers, and the availability of external information such as parental genotypes or population reference panels. Long reads provide phasing information over the distance spanned by a single read, so longer reads generally produce longer haplotype blocks. However, if a region has few heterozygous variants, even long reads may not connect variants across that region, leaving gaps in the phased haplotypes. Switch errors occur when the phasing algorithm assigns a variant to the wrong chromosome copy relative to the surrounding variants. A switch error effectively flips the haplotype assignment for all variants downstream of the error point. The benchmark study of SHAPEIT and Beagle on large cohorts found that switch errors predominantly occur in regions lacking identity-by-descent coverage, which means regions where the individual does not share a long segment with any relative in the dataset [7]. This finding has a direct implication for single-individual phasing: if you have no relatives in your dataset, you have no identity-by-descent information, and your phasing accuracy depends entirely on read length and variant density. Trio-based phasing avoids this problem because the parental genotypes provide independent evidence for every variant. For each variant where the proband is heterozygous and the parents are homozygous for different alleles, the parental origin is unambiguous. For variants where one or both parents are heterozygous, the phase can often be resolved by considering multiple variants together, but some configurations remain ambiguous even with trio data. In practice, trio-based phasing resolves the vast majority of variants, and the remaining ambiguities are usually in regions with low variant density or complex structural variation. ### What Determines Cost The cost of trio-based phasing is not simply three times the cost of single-individual phasing. Library preparation costs scale with the number of samples, so three libraries are needed instead of one. Sequencing costs scale with the amount of data generated, so you need roughly three times the sequencing throughput. Sample collection and DNA extraction costs also apply to three individuals instead of one. If parental samples are not already available, you must account for the logistics of collecting blood or saliva from both parents, which may involve travel, consent, and additional institutional review board approval. There is also a hidden cost in data analysis. Joint calling across three samples requires more computational resources and more careful quality control than analyzing one sample. You need to verify that the parental samples are indeed the biological parents, which requires a relatedness check. You need to handle the possibility of sample mix-ups, which is more likely when more samples are involved. The nf-core documentation describes community standards for reproducible bioinformatics pipelines, and following those standards can reduce the risk of analysis errors but does not eliminate the need for careful sample tracking [5]. ### When Parent-of-Origin Information Is Essential Parent-of-origin effects occur when the phenotypic impact of a variant depends on which parent transmitted it. These effects are important in fetal growth, where a review of genetic association studies identified 38 study-wide significant parent-of-origin effect variants associated with fetal growth traits across 17 studies [8]. The review noted that genome-wide association studies are expanding as computational phasing and parent-of-origin determination methods advance, but candidate gene approaches still dominate the literature [8]. This evidence shows that parent-of-origin analysis is an active research area with clinical relevance, and it requires phasing information that assigns variants to maternal or paternal origin. Trio-based phasing provides parent-of-origin assignment directly from Mendelian inheritance. If the proband is heterozygous and the father is homozygous for allele A while the mother is homozygous for allele B, then the proband's allele A came from the father and allele B came from the mother. Single-individual phasing cannot make this assignment because it has no information about which parent contributed which allele. Some computational methods can infer parent-of-origin using identity-by-descent segments shared with relatives, but this requires having relatives in the dataset and works only in regions covered by those segments [7]. For a single individual with no relatives sequenced, parent-of-origin assignment is not possible. If your research question involves parent-of-origin effects, genomic imprinting, or any analysis where you need to know which parent contributed a variant, trio-based phasing is the appropriate choice. If your research question only requires knowing which variants are on the same chromosome copy, without regard to parental origin, single-individual phasing may be sufficient. ## Practical Workflow for Trio-Based Phasing ### Sample Collection and Verification The first step in a trio-based phasing project is confirming that you have samples from both biological parents. This is not a trivial step. Sample mix-ups, non-paternity, and adoption can all invalidate the trio relationship. You should perform a relatedness check using genotype data from all three samples before proceeding with phasing. This check compares the proportion of the genome shared between each pair of samples and confirms that the proband shares approximately half of their genome with each parent and that the parents share the expected amount for unrelated individuals. The NCBI provides databases and tools for genetic analysis that can support relatedness checks and quality control [1]. The Galaxy Training Network offers accessible tutorials on genomic analysis workflows that include quality control steps, and these tutorials can help laboratory professionals implement reproducible analysis procedures [4]. The Carpentries lessons provide foundational training in shell, Git, and data management that are useful for anyone managing multi-sample sequencing projects [6]. ### Sequencing Strategy for Trios For trio-based phasing with long reads, you need sufficient coverage for all three samples. The proband typically requires the highest coverage because the downstream analysis focuses on their genome. The parents can be sequenced at lower coverage because their role is to provide genotype information for phasing, not to support variant discovery in their own genomes. However, the parents must have enough coverage to call their genotypes accurately at all variant sites in the proband. If a parent has a low-coverage genotype call that is wrong, that error propagates into the phasing of the proband. A common strategy is to sequence the proband at the coverage recommended for your platform and application, then sequence each parent at a coverage level sufficient for accurate genotyping. The exact coverage values depend on your platform, your variant calling pipeline, and the variant types you care about. You should consult the documentation for your sequencing platform and your analysis pipeline to determine appropriate coverage targets. ### Joint Variant Calling and Phasing After sequencing, you should call variants jointly across the three samples instead of calling each sample independently. Joint calling allows the variant caller to use information from all samples to improve genotype accuracy, particularly at sites where one sample has low coverage. The Bioconductor project provides packages for genomic analysis that support joint analysis workflows, and the package documentation includes installation and usage instructions [3]. The phasing step uses the joint genotype calls to assign each variant in the proband to a parental chromosome. Several software tools implement trio phasing, and the choice of tool depends on your data format and analysis pipeline. The key output is a pair of phased haplotypes for the proband, with each variant labeled as maternal or paternal in origin. ### Quality Control for Trio Phasing Quality control for trio phasing should include the following checks: 1. Verify that the Mendelian inheritance patterns are consistent across all three samples. At each variant site, the proband's genotype should be consistent with the parental genotypes under Mendelian inheritance. Sites that violate Mendelian inheritance may indicate genotyping errors, sample mix-ups, or de novo mutations. 2. Check the switch error rate if you have an independent method for validating phasing. This could be a second phasing method, a different sequencing platform, or a known reference from a related individual. 3. Examine the distribution of haplotype block lengths. Trio phasing should produce long blocks, often spanning entire chromosomes in regions with sufficient variant density. Short blocks may indicate low variant density, sequencing errors, or problems with the phasing algorithm. 4. Confirm that the parental origin labels are consistent with the sex of the parents for the X chromosome. The father contributes one X chromosome to daughters and none to sons, so the X chromosome phasing should reflect this pattern. ## Practical Workflow for Single-Individual Phasing ### When Single-Individual Phasing Is Appropriate Single-individual phasing is appropriate when parental samples are unavailable, when the research question does not require parent-of-origin information, or when the budget cannot accommodate three samples. It is also the only option for non-model organisms where parental samples cannot be collected, such as wild-caught individuals or species that are difficult to breed in captivity. The wisent genome assembly project provides a cautionary example of the limitations of single-individual phasing in species with population bottlenecks. The researchers produced a HiFi read-based assembly with a contig N50 of 91 Mb, but they found that extended runs of homozygosity in the wisent genome compromised the separation of HiFi reads into parental-specific read sets, resulting in inferior haplotype assemblies [11]. This evidence shows that single-individual phasing can fail in genomes with long runs of homozygosity, which are common in species that have experienced population bottlenecks or inbreeding. If your study species has low heterozygosity, you should expect phasing challenges and plan accordingly. ### Computational Phasing Methods Single-individual phasing with long reads typically proceeds in two stages. The first stage uses the long reads to phase variants into blocks, where each block contains variants that are connected by reads spanning multiple variants. The second stage may use a reference panel of haplotypes from a population database to order and orient the blocks, or it may leave the blocks unordered if no reference panel is available. The accuracy of the first stage depends on read length and variant density. Longer reads connect more variants, producing longer blocks. Higher variant density means more opportunities for reads to span multiple variants, which also produces longer blocks. In regions with low variant density, the blocks will be short, and the phase between blocks will be unknown. The accuracy of the second stage depends on the reference panel. If the individual is from a population that is well represented in the reference panel, the panel can help order and orient the blocks accurately. If the individual is from a population that is poorly represented, the panel may introduce errors. The benchmark study of phasing methods on large cohorts found that accuracy was high in British samples from the UK Biobank, but the study also noted that errors occur in regions lacking identity-by-descent coverage [7]. This finding suggests that reference panel-based phasing works best when the individual has close relatives in the panel, which is rarely the case for a single individual in a large cohort. ### Quality Control for Single-Individual Phasing Quality control for single-individual phasing should include the following checks: 1. Measure the N50 of your phased blocks. This metric tells you the block length at which half of the phased variants are in blocks of that length or longer. Longer N50 values indicate better phasing. 2. Check the switch error rate if you have a validation method. This could be a second phasing method, a different sequencing platform, or a comparison with a trio-phased individual from the same population. 3. Examine the distribution of block lengths across the genome. Regions with low variant density will have shorter blocks, and this is expected. However, unexpectedly short blocks in high-density regions may indicate sequencing errors or problems with the phasing algorithm. 4. If you are using a reference panel, check the population assignment of your individual. Individuals from populations that are poorly represented in the panel will have lower phasing accuracy. ## Options and Tradeoffs: Choosing Between the Two Approaches ### Research Goals That Favor Trio-Based Phasing Trio-based phasing is the appropriate choice when your research goal requires parent-of-origin information. This includes studies of genomic imprinting, parent-of-origin effects on gene expression, and analyses where you need to distinguish maternal from paternal alleles. The review of parent-of-origin effects in fetal growth identified 38 study-wide significant variants across 17 studies, and the authors noted that genome-wide association studies are expanding as computational phasing and parent-of-origin determination methods advance [8]. If you are planning a study of parent-of-origin effects, you need phasing that assigns variants to parental origin, and trio-based phasing is the most direct way to obtain that information. Trio-based phasing is also appropriate when you need to detect compound heterozygous variants, where both copies of a gene carry different pathogenic variants. To determine whether two variants are on different chromosome copies, you need phasing information. Long-read sequencing can provide this information from a single individual if the variants are close enough to be spanned by a single read, but trio-based phasing provides more reliable phasing across the entire gene. ### Research Goals That Favor Single-Individual Phasing Single-individual phasing is the appropriate choice when parental samples are unavailable or when the research question does not require parent-of-origin information. Population studies, evolutionary analyses, and exploratory projects often fall into this category. If you are sequencing a set of unrelated individuals to characterize genetic diversity in a population, you do not need parent-of-origin information, and single-individual phasing is sufficient. Single-individual phasing is also appropriate when the budget is constrained. Sequencing three samples instead of one triples the sequencing cost, and this may not be justifiable if the research question does not require trio information. The cost difference is substantial, and it should be weighed against the accuracy difference. ### The Middle Ground: Statistical Phasing with Reference Panels There is a middle ground between trio-based and single-individual phasing. If you have a single individual but you also have access to a large reference panel of haplotypes from the same population, you can use statistical phasing methods that leverage the panel to improve accuracy. The benchmark study of SHAPEIT and Beagle on more than eight million samples and the UK Biobank showed that these methods can achieve very low switch error rates in populations that are well represented in the panel [7]. However, the same study found that errors occur in regions lacking identity-by-descent coverage, which means the methods work best when the individual has relatives in the panel [7]. For long-read projects, the practical implication is that you can combine long-read phasing with statistical phasing using a reference panel. The long reads provide local phasing information, and the reference panel provides information about the order and orientation of blocks. This combined approach can produce accurate phasing without parental samples, but it depends on having a suitable reference panel. ### Cost-Benefit Analysis Framework To choose between the two approaches, you should conduct a cost-benefit analysis that considers the following factors: 1. The cost of collecting and sequencing parental samples, including sample collection logistics, DNA extraction, library preparation, and sequencing. 2. The cost of additional data analysis for trio phasing, including joint variant calling, relatedness checks, and quality control. 3. The value of parent-of-origin information for your research question. If your question cannot be answered without parent-of-origin information, trio-based phasing is required regardless of cost. 4. The accuracy requirements of your downstream analysis. If your analysis is sensitive to switch errors, you may need the higher accuracy of trio-based phasing. 5. The availability of a reference panel for your population. If a good reference panel is available, single-individual phasing with statistical methods may achieve sufficient accuracy. ## Records and Measurements for Phasing Projects ### What Records to Keep For any phasing project, you should keep detailed records of the following: 1. Sample identifiers and their relationships. For trio projects, record which samples are the parents and which is the proband. Record the sex of each sample and any relevant clinical or phenotypic information. 2. Sequencing metadata. Record the sequencing platform, the flow cell or chip identifier, the sequencing run date, the coverage achieved for each sample, and the read length distribution. 3. Analysis pipeline versions. Record the versions of all software used for quality control, variant calling, and phasing. This includes the operating system, the pipeline manager, and each individual tool. 4. Quality metrics. Record the quality metrics for each step of the analysis, including read quality scores, mapping rates, variant call quality, and phasing block lengths. 5. Parameter settings. Record the parameters used for each analysis step, including coverage thresholds, quality filters, and phasing algorithm settings. The nf-core documentation describes community standards for reproducible bioinformatics pipelines, and following those standards includes documenting pipeline versions and parameters [5]. The Galaxy Training Network offers tutorials on reproducible analysis workflows that emphasize the importance of recording analysis steps [4]. ### Key Metrics to Monitor The following metrics are important for monitoring phasing quality: 1. Switch error rate. This is the proportion of variants that are assigned to the wrong haplotype relative to the surrounding variants. Lower is better. The benchmark study of phasing methods on large cohorts reported median switch error rates of 0.00 percent in trio children and 0.026 percent in British samples [7]. 2. Haplotype block N50. This is the block length at which half of the phased variants are in blocks of that length or longer. Higher is better. Long-read phasing should produce blocks that are much longer than short-read phasing. 3. Proportion of variants phased. This is the proportion of heterozygous variants that are assigned to a haplotype. Some variants may be unphased if they are in regions with no connecting reads or if they are in low-complexity regions. 4. Mendelian consistency rate for trios. This is the proportion of variant sites where the proband's genotype is consistent with the parental genotypes under Mendelian inheritance. High consistency indicates good genotyping and sample verification. 5. Identity-by-descent coverage for statistical phasing. This is the proportion of the genome covered by identity-by-descent segments shared with relatives in the dataset. The benchmark study found that switch errors predominantly occur in regions lacking identity-by-descent coverage [7]. ### How to Validate Phasing Accuracy Validating phasing accuracy requires an independent source of truth. For trio projects, you can validate the phasing by comparing the phased haplotypes with the parental genotypes. For single-individual projects, you can validate by comparing with a second phasing method, by comparing with a trio-phased individual from the same population, or by using a different sequencing platform. If you have access to a reference panel, you can also validate by checking whether the phased haplotypes are consistent with the haplotypes in the panel. However, this validation is circular if you used the panel for phasing, so it is only meaningful if you used a different method for the initial phasing. ## Common Failure Patterns and How to Address Them ### Failure Pattern 1: Short Haplotype Blocks in Single-Individual Phasing Short haplotype blocks are the most common failure pattern in single-individual phasing. This occurs when the read length is too short to connect multiple variants, when the variant density is too low, or when the phasing algorithm cannot resolve the phase between blocks. The wisent genome assembly project provides an example where extended runs of homozygosity compromised the separation of HiFi reads into parental-specific read sets, resulting in inferior haplotype assemblies [11]. To address this failure pattern, you can increase the read length, increase the sequencing coverage to improve variant detection, or use a reference panel to order and orient the blocks. If the genome has long runs of homozygosity, you may need to accept that phasing will be incomplete in those regions. ### Failure Pattern 2: Mendelian Inconsistencies in Trio Phasing Mendelian inconsistencies occur when the proband's genotype is not consistent with the parental genotypes. This can indicate genotyping errors, sample mix-ups, non-paternity, or de novo mutations. The first step is to verify that the samples are correctly labeled and that the relatedness check confirms the expected relationships. If the samples are correct, the inconsistencies may be due to genotyping errors, which can be addressed by increasing coverage or using a different variant caller. De novo mutations are a legitimate source of Mendelian inconsistencies, and they are of interest in some studies. If you are studying de novo mutations, you should expect a small number of inconsistencies and verify them with an independent method. ### Failure Pattern 3: Switch Errors in Low-Complexity Regions Switch errors in low-complexity regions occur when the phasing algorithm cannot determine the correct order of variants because the region has few informative markers. The benchmark study of phasing methods found that switch errors predominantly occur in regions lacking identity-by-descent coverage [7]. This failure pattern is more common in single-individual phasing than in trio phasing, because trio phasing uses parental genotypes to resolve the phase. To address this failure pattern, you can increase the read length to span more variants, increase the coverage to improve variant detection, or use a reference panel to provide additional information. If the region is truly uninformative, you may need to accept that the phase is unknown in that region. ### Failure Pattern 4: Sample Mix-Ups in Trio Projects Sample mix-ups are a serious risk in trio projects because the phasing depends on the correct assignment of samples to parents and proband. A mix-up can produce completely wrong phasing results that are difficult to detect without careful quality control. The relatedness check is the primary defense against sample mix-ups, and it should be performed before any downstream analysis. To address this failure pattern, you should track samples carefully throughout the workflow, use unique identifiers for each sample, and perform the relatedness check as soon as genotype data are available. If a mix-up is detected, you must determine which samples were affected and re-sequence if necessary. ## Limitations of Both Approaches ### Limitations of Trio-Based Phasing Trio-based phasing has several limitations beyond cost. Parental samples may not be available, particularly for adult probands whose parents are deceased or unavailable. The parents may not be the biological parents, which requires a relatedness check to confirm. The phasing may be incomplete in regions where both parents are heterozygous for the same alleles, because the parental origin cannot be determined from the genotypes alone. Trio-based phasing also assumes that the parents are unrelated. If the parents are related, the phasing becomes more complex because the parental haplotypes may share long segments, reducing the information available for phasing. This is a particular concern in populations with high rates of consanguinity. ### Limitations of Single-Individual Phasing Single-individual phasing has the fundamental limitation that it cannot assign variants to parental origin. This makes it unsuitable for studies of parent-of-origin effects, genomic imprinting, and other analyses that require knowing which parent contributed a variant. The review of parent-of-origin effects in fetal growth identified 38 study-wide significant variants, and the authors noted that most studies were conducted in populations of White European ancestry, indicating a gap in research among more diverse populations [8]. This evidence highlights the importance of parent-of-origin information and the need for diverse study populations. Single-individual phasing also has accuracy limitations in regions with low variant density, long runs of homozygosity, or complex structural variation. The wisent genome assembly project showed that extended runs of homozygosity can compromise the separation of HiFi reads into parental-specific read sets [11]. If your study species or population has low heterozygosity, you should expect phasing challenges. ### Limitations of Reference Panel-Based Phasing Reference panel-based phasing is limited by the representation of the individual's population in the panel. If the individual is from a population that is poorly represented, the phasing accuracy will be lower. The benchmark study of phasing methods found that accuracy was high in British samples from the UK Biobank, but the study also noted that errors occur in regions lacking identity-by-descent coverage [7]. This finding suggests that reference panel-based phasing works best when the individual has close relatives in the panel. Reference panel-based phasing also cannot phase variants that are not in the panel. If your long-read data contain novel variants that are not present in the reference panel, those variants will be unphased or will be phased with lower confidence. ## Welfare and Safety Context for Clinical Applications ### Clinical Diagnostic Context Long-read genome sequencing is increasingly used in clinical diagnostics, particularly for critically ill patients where rapid results can guide treatment decisions. A study of ultrarapid nanopore long-read genome sequencing in critically ill pediatric patients found that a genetic diagnosis was made in 11 of 26 cases, with an average turnaround time of 5.3 days from sample receipt to result, compared with 18.4 days for standard genomic care [9]. The study also found that DNA methylation analysis from long-read sequencing expedited the diagnosis in 3 of 26 cases, and that in 7 of 11 solved cases, the ultrarapid sequencing led to immediate adjustments in patient care [9]. This clinical context has implications for phasing decisions. In a critically ill patient, the speed of diagnosis is paramount, and trio-based phasing may not be feasible if parental samples are not immediately available. However, the clinical utility of the result may depend on knowing whether a variant is inherited or de novo, which requires parental genotypes. The decision to pursue trio-based phasing in a clinical setting must balance the need for speed against the need for parental origin information. ### Barriers to Clinical Implementation A systematic review of genomic sequencing in the neonatal intensive care unit identified several barriers to widespread implementation, including the availability of personnel to facilitate timely diagnosis and coverage of cost [10]. The review also noted that clinicians report worries about a lack of genetics knowledge, informed consent, results return, and potential harm, while parents consistently report low levels of anxiety, decisional conflict, harm, or regret [10]. These findings are relevant to phasing decisions because they highlight the importance of having trained personnel who can interpret phasing results and communicate them to families. The review also found that the lack of availability of translated consent documents limits the participation of families who do not speak English or Spanish [10]. This is a practical consideration for trio-based phasing projects that require parental samples, because the consent process must be accessible to all families. ### Professional Escalation Criteria You should escalate to a specialist when you encounter any of the following situations: 1. You need parent-of-origin information for a clinical diagnosis, and you are considering single-individual phasing because parental samples are unavailable. A genetic counselor or clinical geneticist can help you determine whether the diagnosis can be made without parent-of-origin information. 2. You detect Mendelian inconsistencies in a trio project that cannot be explained by genotyping errors or de novo mutations. This may indicate sample mix-ups or non-paternity, and you should consult a specialist before proceeding. 3. You are working with a species or population with low heterozygosity, and your phasing results are poor. A bioinformatics specialist can help you determine whether alternative phasing strategies are available. 4. You are planning a clinical study that will use phasing results to guide treatment decisions. You should consult with clinical geneticists and bioinformatics specialists to ensure that the phasing accuracy is sufficient for the intended use. 5. You are considering using a reference panel for phasing, and you are unsure whether your population is well represented in the panel. A population geneticist can help you assess the suitability of the panel. ## Practical Implementation Steps ### Step 1: Define Your Research Question and Required Outputs Before choosing a phasing strategy, write down your research question and the specific outputs you need. Do you need parent-of-origin information? Do you need whole-chromosome haplotypes, or are local haplotypes sufficient? What level of switch error is acceptable for your downstream analysis? Answering these questions will narrow your options. ### Step 2: Assess Sample Availability and Budget Determine whether parental samples are available and whether the budget can accommodate sequencing three samples instead of one. If parental samples are not available, single-individual phasing is your only option. If parental samples are available but the budget is constrained, you may need to choose between trio-based phasing for a smaller number of samples and single-individual phasing for a larger number. ### Step 3: Evaluate Reference Panel Availability If you are considering single-individual phasing, determine whether a suitable reference panel is available for your population. The panel should be large enough to provide accurate phasing and should include individuals from the same population as your sample. If no suitable panel is available, you will need to rely on long-read phasing alone, which may produce shorter blocks. ### Step 4: Choose Your Sequencing Strategy Based on your research question, sample availability, budget, and reference panel availability, choose your sequencing strategy. For trio-based phasing, plan to sequence the proband at the coverage recommended for your application and the parents at coverage sufficient for accurate genotyping. For single-individual phasing, plan to sequence the individual at the coverage recommended for your application and consider whether you need to generate additional data for phasing. ### Step 5: Implement Quality Control Checks Implement the quality control checks described in this article. For trio projects, perform a relatedness check before proceeding with phasing. For all projects, track the key metrics described in the records and measurements section and document your analysis pipeline versions and parameters. ### Step 6: Document Your Decision and Rationale Document your decision to use trio-based or single-individual phasing, including the rationale based on your research question, sample availability, budget, and reference panel availability. This documentation will be useful for publications, grant reports, and future projects. ## Frequently Asked Questions ### What is the main advantage of trio-based phasing over single-individual phasing? The main advantage of trio-based phasing is that it provides direct experimental evidence for which variants are on which chromosome copy and which parent contributed each variant. This parent-of-origin information is essential for studies of genomic imprinting and parent-of-origin effects, and it cannot be obtained from single-individual phasing without additional relatives or inference methods. Trio-based phasing also produces longer haplotype blocks and is less affected by low variant density or long runs of homozygosity. ### Can single-individual phasing achieve the same accuracy as trio-based phasing? Single-individual phasing can achieve high accuracy in some contexts, particularly when a large reference panel is available and the individual is from a population that is well represented in the panel. A benchmark study of phasing methods on more than eight million samples found median switch error rates of 0.00 percent in trio children and 0.026 percent in British samples [7]. However, the same study found that errors occur in regions lacking identity-by-descent coverage, and single-individual phasing cannot assign variants to parental origin [7]. ### When should I use a reference panel for phasing? You should use a reference panel for phasing when you have a single individual, no parental samples, and a reference panel that includes individuals from the same population as your sample. The panel can help order and orient haplotype blocks and improve accuracy in regions where long reads do not provide sufficient information. However, you should be aware that the panel may introduce errors if your population is poorly represented. ### What is a switch error rate and why does it matter? A switch error rate measures the proportion of variants that are assigned to the wrong haplotype relative to the surrounding variants. A switch error effectively flips the haplotype assignment for all variants downstream of the error point, which can have a large impact on downstream analyses. Lower switch error rates indicate more accurate phasing. The benchmark study of phasing methods reported median switch error rates of 0.00 percent in trio children and 0.026 percent in British samples [7]. ### How do I know if my phasing results are accurate enough for my analysis? You can assess phasing accuracy by measuring the switch error rate if you have a validation method, by examining the haplotype block N50, and by checking the proportion of variants phased. The accuracy requirements depend on your downstream analysis. Analyses that are sensitive to switch errors, such as detecting compound heterozygous variants or studying allele-specific expression, require higher accuracy than analyses that are less sensitive. ### What should I do if my phasing results show short haplotype blocks? Short haplotype blocks can result from low variant density, short reads, or long runs of homozygosity. You can try increasing the read length, increasing the sequencing coverage, or using a reference panel to order and orient the blocks. If the genome has long runs of homozygosity, you may need to accept that phasing will be incomplete in those regions. The wisent genome assembly project provides an example where extended runs of homozygosity compromised haplotype assembly [11]. ### Is trio-based phasing worth the extra cost? Trio-based phasing is worth the extra cost when you need parent-of-origin information, when your analysis is sensitive to switch errors, or when you need whole-chromosome haplotypes. The cost is approximately three times the single-genome cost, and you must also account for the logistics of collecting parental samples. If your research question does not require parent-of-origin information and your analysis is not sensitive to switch errors, single-individual phasing may be sufficient. ### What are the limitations of phasing in non-model organisms? Phasing in non-model organisms is limited by the availability of reference panels, the heterozygosity of the genome, and the availability of parental samples. The wisent genome assembly project showed that extended runs of homozygosity in a species that went through a population bottleneck compromised the separation of HiFi reads into parental-specific read sets [11]. If your study species has low heterozygosity, you should expect phasing challenges and plan accordingly. ## Related Bioinformatics Guides - [Long-Read Sequencing Cost and Market: What to Expect](/knowledge/bioinformatics/long-read-sequencing-cost-and-market-what-to-expect) - [How to Choose a Long-Read Sequencing Platform: PacBio vs Oxford Nanopore](/knowledge/bioinformatics/how-to-choose-a-long-read-sequencing-platform-pacbio-vs-oxford-nanopore) - [Detecting Structural Variants with Long-Read Sequencing: Methods and Considerations](/knowledge/bioinformatics/detecting-structural-variants-with-long-read-sequencing-methods-and-considerations) - [Long-Read Sequencing for De Novo Assembly of Complex Genomes: Case Studies and Best Practices](/knowledge/bioinformatics/long-read-sequencing-for-de-novo-assembly-of-complex-genomes-case-studies-and-best-practices) - [Single-Cell Sequencing Services: How to Choose a Provider](/knowledge/bioinformatics/single-cell-sequencing-services-how-to-choose-a-provider) ## References and Further Reading - [NCBI Data Resources](https://www.ncbi.nlm.nih.gov/). National Center for Biotechnology Information. - [EMBL-EBI Training](https://www.ebi.ac.uk/training). European Bioinformatics Institute. - [Bioconductor](https://bioconductor.org/). Bioconductor Project. - [Galaxy Training Network](https://training.galaxyproject.org/). Galaxy Project. - [nf-core Documentation](https://nf-co.re/docs). nf-core. - [The Carpentries Lessons](https://carpentries.org/lessons). The Carpentries. - [Phasing millions of samples achieves near perfect accuracy, enabling parent-of-origin analyses.](https://doi.org/10.1016/j.xhgg.2025.100479). 2025. - [A rapid review of genetic association studies of parent-of-origin effects and fetal growth.](https://doi.org/10.1186/s40348-026-00224-5). 2026. - [Nanopore long-read sequencing for the critically ill facilitates ultrarapid diagnostics and urgent clinical decision making.](https://doi.org/10.1038/s41431-025-01959-x). 2026. - [Benefits and barriers to broad implementation of genomic sequencing in the NICU.](https://doi.org/10.1016/j.ajhg.2025.04.007). 2025. - [Genome assembly of wisent (Bison bonasus) uncovers a deletion that likely inactivates the THRSP gene.](https://doi.org/10.1038/s42003-024-07295-y). 2024. > This article is educational and does not replace validated analysis plans, institutional policy, clinical interpretation, or specialist review.