Effective use of data drives efficient, reliable analysis in next-generation sequencing. Users of the VarSeq suite are well-acquainted with the tools that VarSeq supplies to empower users, at the center of which lies a curated collection of clinical annotations that users can leverage in their own customized workflows. In the decades that we’ve improved and expanded the VarSeq suite since its release, we’ve also tracked the moving target of clinical knowledge. In 2026, users can explore 385 individual annotation tracks, consisting of 13.1 billion records — 22.5% of which weren’t in our database 5 years ago. More isn’t always better though; that’s why we employ a comprehensive curation process each month to ensure that what we’re shipping is consistent and applicable, and that new additions can be implemented into user workflows confidently, while giving users the autonomy to decide when and which tracks to update. Let’s break down some of the most poignant tracks in our library, and explore how they’ve improved not just in quantity but in quality over time, with more actionable records at our users’ fingertips.

What we curate

At a glance, our data source library, consisting of both open-source and proprietary tracks, consists of a growing list of consortia, registries, and peer-reviewed sources, with clinical mainstays like ClinVar, ClinGen, and CIViC, industry-standard gene tracks like OMIM, and specialized tracks like BRCA Exchange. In addition, we provide our own in-house, expert-curated Cancer Knowledge Base, which is the subject of previous and future detailed blog posts.

Different tracks come from different places and serve different purposes, and thus we see appropriate growth curves for individual tracks (Fig. 1). While knowledge bases like ClinVar represent the near-exponential growth of amalgamated industry expertise, other tracks like OMIM and ClinGen reflect growth not necessarily in the number of records, but in the quality of those records as information and actionability is augmented over time (Fig. 5).

Small or large, we stay abreast of all changes made to tracks during our curation process. Smaller tracks may have fewer records, but often the individual records may prove more salient to our users investigating relevant issues (Fig. 2).

A window into clinical knowledge

As we mentioned above, quantity doesn’t necessarily imply quality. The actionability of new records is the indicator of value. In the case of ClinVar, we can see that, while the total number of records has increased substantially, the number of benign, likely benign, pathogenic, and likely pathogenic variants have kept abreast of VUS submissions (Fig. 3). Each new submission adds evidence to user workflows that can be used to minimize the number of false positives on the benign side, and ensure capture of true positives in the pathogenic case. In the case of CIViC, we see a similar story: over time, the number of Level A — Validated and Level B — Clinical increases at pace with the overall number of records, even outpacing lower-impact submissions over the past few years (Fig. 4).

ClinGen Gene Dosage tells a similar story on the haploinsufficiency front. As the number of records increase, we see a higher ration of actionable haploinsufficiency and autosomal recessive submissions, decreasing the relative grey area of insufficient evidence (Fig. 5).

Bringing it all together

Naturally, we expect all of these results: legions of hard-working clinicians and researchers around the globe contribute to these tracks, so it’s no surprise that their utility increases over time. With the VarSeq suite, paired with our comprehensive curation efforts, however, users can access all of this new data automatically and seamlessly, without relying on their own curation process or manually querying records. Instead, existing workflows are updated when the user chooses to incorporate new track versions.

Questions, comments, or commendations for this post? Or is your workflow in need of an upgrade that allows you to automatically incorporate the vast amount of clinical knowledge available in our 300+ curated annotation tracks? Reach out to our team, we’re looking forward to hearing from you!

The post A Decade of Clinical Genomics in Curated Annotations appeared first on The Golden Helix Blog.

DNA methylation plays a critical role in gene regulation and is increasingly recognized as a clinically meaningful biomarker in oncology. With the rise of long-read sequencing, we now have the ability to detect base modifications alongside small variants and structural variations. In this post, we’ll walk through how VSWarehouse and VarSeq support methylation analysis, from raw PacBio reads all the way to clinical reporting.

A Somatic Workflow Built for Long Reads

VSWarehouse provides a PacBio WGS Somatic Workflow, available for download from our Workflow Repository. This automated pipeline handles everything from alignment to variant calling, phasing, structural variant detection, and methylation analysis, all from a single sequencing run.

At the heart of our methylation support is the CpG Pileup and DMR Analysis stage, which performs two key tasks: generating per-CpG methylation scores across the genome, and identifying differentially methylated regions (DMRs) between tumor and matched normal samples.

Identifying Differentially Methylated Regions

The DMR output captures the genomic coordinates of each region along with the number of CpG sites covered, mean methylation levels in the tumor and normal, the difference between them, and a statistical area score reflecting the strength of the signal. The mean methylation differences between tumor and normal tissue can point to biologically and potentially clinically significant epigenetic alterations.

DMRs can be imported directly into VarSeq, where they appear as a filterable table allowing users to sort and filter on methylation difference, region length, statistical significance, and genomic location, making it straightforward to prioritize the most meaningful epigenetic events.

Visualizing Methylation Data

VarSeq’s integrated genome browser can render base modification data stored directly in long-read BAM files (one of the unique advantages of PacBio HiFi sequencing over short-read technologies). The coloring scheme is designed for immediate visual clarity: bases with a high probability of methylation are colored red, while bases with a low methylation probability are colored blue. This intuitive display makes it easy to spot regions of aberrant hypermethylation at single-base resolution.

Reporting Methylation Biomarkers in VSClinical

While our variant interpretation and reporting platform, VSClinical, is primarily oriented toward sequence variants, copy number variants, and structural variations, we recognize that epigenetic biomarkers are increasingly important in clinical oncology. A well-established example is MGMT promoter methylation in glioblastoma, which is both a prognostic marker and a predictor of response to temozolomide chemotherapy.

To support these use cases, VSClinical allows users to create and report user-defined interpretations for methylated DNA biomarkers, enabling labs to document and communicate clinically relevant methylation findings within their standard reporting workflow.

Conclusion

Long-read sequencing technology provides the unique ability to analyze both genetic variation and epigenetic modifications using the same dataset. Through our PacBio WGS Somatic Workflow, we are able to deliver a truly integrated view of the cancer genome, including sequence variants, structural variants, phasing, and epigenetics, all from a single sequencing run. If you’re interested in learning more about how VarSeq and VSWarehouse can support your long-read sequencing workflows, please reach out to our team using the contact link below.

The post Long-Read Part 2: Methylation Analysis appeared first on The Golden Helix Blog.

If you spend enough time staring at VCF files, you eventually start to feel like you’re reading a secret language. Most of the time, things are straightforward with the Genotype (GT) representation: 0/0, 0/1, 1/1. Easy enough. But then one day, you import your VCFs for a trio analysis and see GT fields with ./., 0/., ./1, or even 1|0 staring back at you.

At first glance, it can feel like the VCF is trying to mess with you. In reality, these genotype representations are perfectly valid and they may show up when you’re working with family analysis such as a Trio or a large cohort that you often see with research projects. Fortunately, VarSeq is designed to handle these scenarios gracefully. Let’s decode a few of the usual GT suspects.

The Case of the Missing Genotype: ./.

The GT of ./. simply means that no genotype call was assigned for either allele in that sample. This representation often appears in cohort or trio analysis where a variant is present in one individual but not confidently detected in others. Rather than removing the variant entirely, the VCF keeps the record so the position can still be evaluated across all samples.

VarSeq preserves this structure during import so you can compare variants across the entire cohort, even when some samples don’t carry the variant. This is particularly helpful when searching for inheritance patterns in family analyses or exploring rare variants across large groups.

The Half-Missing Genotypes: 0/. and ./1

Sometimes you’ll see a GT where only one allele seems defined. For example:

• 0/. means the reference allele present and the second allele missing
• ./1 means that the alternate allele is present but the reference allele missing

These cases often appear when multi-allelic variants are split into multiple records during processing. The original site may have contained several alternate alleles, and when those are separated into individual rows, the GT representation can look a little unusual.

It may look strange at first, but it’s usually just a reflection of how the variant record was parsed, not an indication that something is wrong with the data. VarSeq has several ways of dealing with complex allele calls, which are covered here in this webinar.

Why Some Samples Have Depth and Others Don’t

Another common observation: one sample has depth metrics and allele counts, while another sample just shows missing values.

This happens because the variant may only have been called in one sample. When VarSeq imports a cohort VCF, it still lists the variant across all samples so you can analyze it collectively. The sample with the variant will have metrics like DP, GQ, or VAF, while the others will show missing values. In other words, the variant exists in the dataset, but not necessarily in every individual.

PASS… For Some Samples

You may also notice that not every sample meets the PASS filter for a given variant. This is often the result of single-sample variant calling pipelines where each sample is processed independently. When those results are combined into a cohort VCF, filtering outcomes can vary slightly from sample to sample.

If you prefer more consistent GT representation across a cohort or trio, joint calling can help. Joint calling analyzes all samples together, improving genotype consistency and often reducing missing calls.

Phasing: When the Slash Turns Into a Pipe

Occasionally the GT delimiter changes from a slash to a pipe: 1|0 instead of 0/1. That little vertical bar indicates phasing. In other words, the software knows which allele belongs to which chromosome copy. Phasing information is increasingly common in long-read sequencing datasets.

This is where VarSeq really shines. Using fields like Phased Set (PS), VarSeq can evaluate whether variants occur in cis or in trans, enabling powerful workflows such as compound heterozygous analysis in trios. Read more about Phasing in VarSeq here.

When in Doubt, Let VarSeq (and the FAS Team) Help

Large datasets, family analyses, and multi-allelic variants can produce genotype fields that look a little unconventional. But these representations are usually just the natural result of combining complex genomic data across multiple samples.

The good news is that VarSeq was built for exactly this kind of analysis, whether you’re working with trios, cohorts, or large sequencing studies.

And if a genotype still looks mysterious? The Golden Helix FAS team is always happy to help interpret what you’re seeing and make sure your analysis pipeline is working exactly the way it should. Because when the VCF gets weird, it’s nice to have experts, and great software, on your side. Contact support@goldenhelix.com for more information today!

The post Genotype: Unusual GT Fields in Cohort and Trio Analysis appeared first on The Golden Helix Blog.

On International CdLS Awareness Day, we recognize the individuals and families affected by Cornelia de Lange Syndrome (CdLS), a rare developmental condition that highlights both the complexity of human genetics and the importance of accurate, timely diagnosis.

CdLS is a genetic disorder typically caused by de novo mutations, meaning the change in DNA occurs early in development rather than being inherited from a parent. It is classified as a cohesinopathy, involving genes that regulate the cohesin complex, an essential system for chromosome structure, gene expression, and normal development.

A Spectrum Rooted in Genetics

CdLS presents as a spectrum, with a wide range of clinical features that can include differences in growth, limb development, cognition, and organ systems. Because of this variability, diagnosis can sometimes be delayed, particularly in milder cases.

The condition is estimated to occur in approximately 1 in 10,000 live births, though this is likely an underestimate due to underdiagnosis (CdLS Foundation). The most commonly implicated gene is NIPBL, with additional contributions from genes such as SMC1A, SMC3, RAD21, and HDAC8 (MedlinePlus Genetics, 2024). While genetic testing identifies a causative variant in about 80% of cases, making a strong argument for genetic screening, clinical expertise remains essential for an accurate diagnosis.

Why De Novo Detection Matters

Because CdLS is most often caused by de novo variants, identifying the causative mutation requires careful comparison between a child and their parents. This is where trio-based genomic analysis becomes especially powerful. By analyzing sequencing data from the affected individual alongside both parents, clinicians can:

• Pinpoint new (de novo) variants
• Filter out inherited background variants
• Prioritize variants in genes known to impact development

This approach has become a cornerstone of rare disease diagnostics, significantly increasing diagnostic yield compared to single-sample analysis (Wright et al., Nature Reviews Genetics, 2018; Stark et al., Genetics in Medicine, 2019).

Supporting Rare Disease Diagnosis with VarSeq

Tools like VarSeq are widely used to support trio analysis workflows, helping labs efficiently identify de novo variants that may underlie conditions like CdLS. With VarSeq, users can:

• Apply inheritance-based filtering to rapidly isolate candidate de novo variants
• Prioritize variants in clinically relevant genes such as NIPBL and related cohesin genes
• Integrate phenotype-driven analysis to connect genomic findings with clinical presentation

In conditions where the causal variant may be rare, subtle, or unexpected, these capabilities can help reduce the time to diagnosis, providing clarity for clinicians and, importantly, answers for families.

Moving Forward with Awareness and Precision

For families affected by CdLS, a diagnosis is more than a label, it’s a pathway to understanding, care, and community. Continued awareness, combined with advances in genomic analysis, is helping make that path clearer.

On this International CdLS Awareness Day, we at Golden Helix recognize the importance of early detection, accurate genetic interpretation, and compassionate clinical care, and the role that modern genomic tools play in supporting all three. For more information about how VarSeq can be used in rare disease detection, please see our Webinar on single sample and family based genome analysis, or reach out to us a info@goldenhelix.com.

Further Resources for Cornelia de Lange Syndrome

U.S. National Library of Medicine. (2022, April 13). Cornelia de Lange Syndrome: Medlineplus genetics. MedlinePlus. https://medlineplus.gov/genetics/condition/cornelia-de-lange-syndrome/

What is CdLS?. CdLS Foundation. (2025, June 12). https://www.cdlsusa.org/what-is-cdls/

Wright, C. F., FitzPatrick, D. R., & Firth, H. V. (2018). Paediatric genomics: Diagnosing rare disease in children. Nature Reviews Genetics, 19(5), 253–268. https://doi.org/10.1038/nrg.2017.116

Stark, Z., Dolman, L., Manolio, T. A., Ozenberger, B., Hill, S. L., Caulfied, M. J., Levy, Y., Glazer, D., Wilson, J., Lawler, M., Boughtwood, T., Braithwaite, J., Goodhand, P., Birney, E., & North, K. N. (2019). Integrating genomics into healthcare: A global responsibility. The American Journal of Human Genetics, 104(1), 13–20. https://doi.org/10.1016/j.ajhg.2018.11.014

The post Understanding Cornelia de Lange Syndrome: The Role of Genomics in Finding Answers appeared first on The Golden Helix Blog.

This May, we’re recognizing #GrayMay, a month dedicated to raising awareness about brain tumors. In this blog post, I want to share something practical regarding precision medicine of gliomas: for gliomas, the diagnosis is no longer based on histology alone. Molecular profiling now helps define the tumor entity, clarify prognosis, and increasingly guide treatment options. If your lab supports CNS cases, molecular profiling can increase diagnostic yield.

Glioma diagnosis is increasingly defined by molecular markers

Two results often set the direction immediately:

• IDH mutation status is a major branching point in diffuse gliomas.
• 1p/19q codeletion is classification-defining when paired with an IDH mutation.

After that, a small set of additional markers does an outsized amount of clinical work. Depending on context, they can support classification, refine prognosis, or indicate glioblastoma-like biology even when histology is ambiguous. Common examples include ATRX, TP53, TERT promoter, EGFR amplification, +7/−10, and CDKN2A/B alterations. Many labs also capture MGMT promoter methylation as part of the clinical picture.

It’s not just small variants anymore

Glioma profiling is multi-variant-type by nature. A practical workflow usually needs to handle:

• Small variants like IDH1/2, TERT promoter, TP53
• Copy-number changes like 1p/19q codeletion, EGFR amplification, CDKN2A/B loss
• Structural variants and fusions, including actionable events like NTRK fusions and glioma-relevant rearrangements such as FGFR3–TACC3
• Pediatric and young adult markers such as H3 K27 / H3 G34 alterations
• Select actionable alterations like BRAF V600E in the right clinical setting

This is where many teams get stuck operationally. The data exists, but the evidence is scattered across spreadsheets, PDFs, and one-off notes, which makes consistency hard across reviewers and across time.

Enterprise genomics is an operations problem

Even strong labs hit the same bottleneck: running the assay is not the hard part. The hard part is doing it the same way every time, across teams, across time, with clean traceability and a reliable handoff into interpretation and reporting.

That is the core value of VSWarehouse 3 for neuro-oncology programs. It is not just a data store. It is the enterprise platform and deployment engine that lets teams run the VarSeq Suite at scale, on-premises or in a private cloud, with the controls institutions expect.

In a CNS workflow, that translates to a few concrete advantages:

• Deploy and execute: run pipelines under a governed system so you can answer what ran, when, with what inputs, and where the logs live.
• Annotate from history: reuse your own prior observations and internal frequencies during new analyses, so interpretation gets faster and more consistent over time.
• Assessment catalogs: codify variant classifications and interpretive notes into a shared knowledge base so new cases benefit from prior decisions.
• Enterprise security: support single sign-on, role-based access control, and isolated workspaces so the platform fits institutional requirements.
• Integrate and exchange: connect to upstream and downstream systems through a REST API, including laboratory and clinical systems, and standardize outputs for reporting and follow-up.

For gliomas, the message is simple: if molecular markers are required for classification, then the workflow needs enterprise-grade repeatability, governance, and reuse of institutional knowledge, not just a way to open a variant file.

Brain tumors by the numbers

The following quick stats come from the National Brain Tumor Society and are useful for #GrayMay awareness posts:

• 93,000+ Americans will be diagnosed with a primary brain tumor in 2025
• 35.7% is the five-year relative survival rate for patients with malignant brain tumors
• 18,330 Americans will die from a malignant brain tumor in 2025

Genomics is how we get to the right glioma label and the right pathway. The enterprise challenge is making that repeatable: consistent evidence capture, consistent interpretation, and consistent reporting, at the scale your program is expected to run.

The post Brain Tumor Awareness Month appeared first on The Golden Helix Blog.

Each month, we highlight new research from the scientific community that advances our understanding of complex genetic diseases and showcases the tools researchers rely on for precise variant interpretation. Our customer publications in April explored healthy aging genes in Super Seniors, chronic kidney disease, and breast cancer risk stratification.

Title: Genome-Wide Association Study and Pathway Analysis of Healthy Aging in Super Seniors

Background: Genome-wide association studies have identified some genetic loci linked to longevity and health span, but only a few, such as APOE and FOXO3, are consistently replicated across populations. This inconsistency likely reflects the significant role of environmental and lifestyle factors, as well as differences in how healthy aging is defined. While genetics do contribute to lifespan, they explain only about 15–40% of the variation in human longevity.

Objective: This article aims to better understand resilience to age-related diseases in Super Seniors, including cancer, cardiovascular disease, diabetes, pulmonary disease, and dementia. This exceptionally healthy group offers a valuable opportunity to identify genetic factors that may protect against disease and support healthy aging.

Subjects and Methods: Researchers collected and genotyped DNA samples from 1017 individuals (597 cases and 420 controls), applied extensive quality control, and focused on individuals of European ancestry, yielding millions of high-quality genetic variants for analysis. They then conducted targeted analyses of known longevity-related genes (like APOE and FOXO3) and a genome-wide association study (GWAS) to identify genetic factors linked to healthy aging, including potential sex-specific effects

Results: This study expanded earlier analyses of longevity-related genes in a larger sample, confirming that variants in APOE, particularly the ε4 type, are associated with lower odds of healthy aging, with stronger effects observed in females. Additional variants near APOE (in TOMM40 and APOC1) and in FOXO3 showed modest or sex-specific associations, further supporting their role in aging and disease resistance. Overall, the results reinforce APOE ε4 as a risk factor and suggest FOXO3 contributes to healthy aging, especially in women. A genome-wide analysis also indicated that healthy aging is likely influenced by many genetic variants with small effects, rather than a few major ones.

Conclusions: Healthy aging appears to have a polygenic basis, meaning it is influenced by many genetic variants with small effects acting through interconnected biological pathways. These pathways—such as those involved in immune function, metabolism, stress response, and cell survival—work together to support longevity and health. Further functional studies are needed to clarify these mechanisms and to help develop interventions that could promote healthy aging even in individuals without a strong genetic advantage.

How SVS was used: “A total of 1127 DNA samples (673 Super Seniors and 454 controls) were genotyped for 4,559,465 single nucleotide polymorphisms (SNPs) using a custom Infinium Omni5Exome-4 v1.3 Bead Chip (Illumina, San Diego, CA, USA) at the McGill University/Genome Quebec Innovation Centre. Quality control (QC) was conducted using Golden Helix and PLINK v1.9.0-b7.7 [16]. QC procedures included re-clustering, removal of duplicate and tri-allelic SNPs, strand error checks, and assessments of sex discrepancies (8 samples removed) and relatedness. Related individuals were identified using KING v2.2.8 [17] with a 3rd-degree relatedness threshold, yielding 39 related pairs. Of these, 24 pairs were replicated samples included for QC purposes and 15 represented biological relatedness of participants; in the latter case, the younger individual was removed (39 samples in total). Genetic principal component analysis (PCA) was used to filter to include only the largest ethnicity group, individuals of European ancestry (44 individuals removed). In the final step, individuals exhibiting excessive heterozygosity (± 3 SD) were removed (16 individuals removed). Afterward, 1020 unrelated individuals of European ancestry (599 Super Seniors and 421 controls) and 3,482,546 variants remained.”

Citation: Hoque, R., Leach, S. & Brooks-Wilson, A. Genome-wide association study and pathway analysis of healthy aging in Super Seniors. GeroScience (2026). https://doi.org/10.1007/s11357-026-02229-4

Title: A Whole-Exome Sequencing-Based Exploration of Chronic Kidney Disease of Unknown Etiology (CKDu) in an Endemic Population in Sri Lanka

Background: Chronic kidney disease (CKD) is a growing global health problem, driven largely by conditions like diabetes, hypertension, and aging, and influenced by both genetic and environmental factors. It disproportionately affects disadvantaged populations and can lead to major financial and health burdens, especially in advanced stages requiring dialysis or transplantation.

A related condition, CKDu (chronic kidney disease of unknown cause), has emerged in agricultural communities in tropical regions, where environmental exposures such as heat stress, toxins, and agrochemicals are suspected contributors. However, patterns of familial clustering suggest that genetic susceptibility may also play a role, even though the exact causes remain unclear.

Objective: Whole-exome sequencing studies of CKDu are currently limited, highlighting the need for larger, more comprehensive genomic research that includes diverse populations and examines interactions between genetic and environmental factors. This study aims to address that gap by analyzing patients and control groups from both endemic and non-endemic regions in Sri Lanka to better understand the genetic basis of the disease.

Subjects and Methods: This study recruited CKDu patients from a high-risk region and compared them with healthy individuals from both endemic and non-endemic areas, ensuring that all participants were carefully screened to exclude known risk factors such as diabetes and hypertension. Blood samples were collected, DNA was extracted, and whole-exome sequencing was performed, followed by rigorous quality control and filtering to identify rare, potentially disease-related genetic variants.

The analysis focused on variants likely to impact gene function and linked them to kidney-related traits using established databases and computational tools, enabling comparisons between patients and controls to identify genetic factors associated with CKDu.

Results: Whole-exome sequencing of 86 participants identified over 4 million genetic variants, which were progressively filtered to focus on rare, high-confidence variants likely to affect protein function and be relevant to CKDu. This process narrowed the dataset to 173 potentially pathogenic variants across 121 genes مرتبط with kidney function and toxin handling.

Several genes, particularly LFNG, PNLDC1, and ATXN3, showed higher variant prevalence in CKDu patients, with LFNG demonstrating a statistically significant association. Overall, CKDu individuals were more likely to carry multiple co-occurring variants, supporting the idea that the disease may result from the combined effects of multiple genetic factors rather than a single cause.

Conclusions: This study shows that chronic kidney disease of unknown etiology (CKDu) in Sri Lanka likely arises from a complex mix of genetic and environmental factors. Although ATXN3 variants were common, they don’t appear to play a major direct role in causing the disease. In contrast, LFNG variants may contribute to CKDu by disrupting kidney repair mechanisms, possibly through altered Notch signaling. The mixed pattern of HLA-DRB1 variants further suggests that disease risk depends on interactions among genetic, environmental, and immune factors.

How VarSeq v2.6.2 was used: “Exome libraries were constructed using Illumina’s DNA prep with exome 2.5 enrichment, per the manufacturer’s instructions (Illumina Inc., San Diego, CA, USA, catalog #: 20077595). Libraries were sequenced on the Illumina NovaSeqX platform, 150 bp paired-end reads to an average read depth of 84,475,876 reads per sample. Raw reads underwent QC, mapping and variant calling using Sentieon’s DNAscope pipeline with the GRCh38 human reference genome build (Sentieon Inc., San Jose, CA, USA). Variant calls from DNAscope for each sample were used as the input for interpretation using Golden Helix’s VarSeq v2.6.2 (Golden Helix, Inc., Bozeman, MT, USA, www.goldenhelix.com) variant filtration and interpretation software. In total, 74 endemic samples and 12 non-endemic samples in the cohort were sequenced (n = 86). Variants were filtered first for quality and had to have an average read depth of 50, a genotype quality Q-score had to be ≥20, and the variant allele fraction had to be ≥0.1.”

Citation: Tom, W., Weerakoon, C., Fernando, N., Hasantha, I., Bandara, M., Krzyzanowski, G., Nanayakkara, S., Cosgrove, D., Nanayakkara, N., & Fernando, M. R. (2026). A Whole-Exome Sequencing-Based Exploration of Chronic Kidney Disease of Unknown Etiology (CKDu) in an Endemic Population in Sri Lanka. International Journal of Molecular Sciences, 27(8), 3369. https://doi.org/10.3390/ijms27083369

Title: Integrating Baseline ctDNA-Derived Tumor Metrics Enhances Risk Stratification in HR-Positive/HER2-Negative Advanced Breast Cancer: A Real-World Multicenter Cohort Study from Austria

Background: Advances in endocrine therapies have significantly improved outcomes for patients with HR-positive/HER2-negative advanced breast cancer. However, a major challenge remains in effectively using molecular and clinical risk factors to guide personalized treatment decisions.

Objective: This article aimed to use circulating tumor DNA (ctDNA) to analyze the genomic and prognostic features of patients with HR-positive/HER2-negative advanced breast cancer in early treatment lines. By taking a tumor-agnostic approach, the researchers aimed to more meaningfully stratify patients using baseline ctDNA metrics.

Subjects and Methods: This prospective study enrolled patients with HR-positive/HER2-negative advanced breast cancer across three Austrian centers and collected blood samples at diagnosis or treatment progression. Researchers analyzed circulating tumor DNA (ctDNA) from plasma using next-generation sequencing to identify genetic alterations and estimate tumor burden through multiple complementary methods. They applied rigorous variant filtering and statistical analyses to assess genomic patterns and their association with patient outcomes. Overall, the study aimed to evaluate the prognostic value of ctDNA metrics and improve risk stratification in real-world clinical settings.

Results: The study found that most blood samples from patients with advanced breast cancer, analyzed, contained detectable genetic mutations. Mutations in genes such as PIK3CA, TP53, and ESR1 were most common, with mutation frequency and complexity increasing in later treatment lines—especially for ESR1, which was often newly acquired after first-line therapy.

Measures of circulating tumor DNA (ctDNA) showed higher tumor burden in later-line patients and were strongly associated with worse clinical outcomes. Importantly, combining multiple ctDNA metrics improved detection and risk stratification, with higher ctDNA levels consistently predicting shorter progression-free and overall survival.

Conclusions: The findings show that ctDNA-based metrics have strong prognostic value in HR-positive/HER2-negative advanced breast cancer. Combining multiple ctDNA measures improves risk stratification beyond traditional clinical factors, making it a useful tool for baseline assessment. Future studies incorporating both baseline and ongoing ctDNA monitoring may further enhance personalized, response-adapted treatment strategies.

How VarSeq v2.2.0 was used: “Variant calling was carried out using the AVENIO Oncology Analysis Software (version 2.1.0, Roche, Basel, Switzerland), with customized filtration settings: variants with a minor allele frequency ≥1% as defined by ExAC version 1.0 or 1000 Genomes version phase_3_v5b databases, or those listed as common single nucleotide polymorphisms (SNPs) in the dbSNP150 database, were excluded by the software. Putative germline variants (characterized by a VAF ∼50% but a low TFx context) were additionally removed. To enable a high confidence variant call set, variants that passed these filters but had <10 mutated reads or a VAF below the assay limit of detection (LOD), as well as recurrent low-level variants observed in multiple patients (suggesting a sequencing or assay artifact) were flagged and manually excluded. The remaining variants were annotated and classified according to their pathogenicity using Golden Helix VarSeq v2.2.0 (Golden Helix Inc., Bozeman, MT) and the OncoKB database.²⁵ For the primary analyses, a VAF threshold of 0.5% was applied to prioritize specificity and analytical confidence over maximal sensitivity, given the risk of false-positive variant calls from the hematopoietic background. Additionally, a secondary analysis using a lower detection threshold (LOD = 0.1% VAF) was conducted to assess the impact of including low-VAF variants on overall genomic landscape patterns.”

Citation: N. Dobrić, S.O. Hasenleithner, C. Suppan, E.V. Klocker, D. Hlauschek, R. Graf, C. Beichler, C. Albertini, D. Egle, D. Liu, A.M. Starzer, R. Bartsch, T. Moser, G. Rinnerthaler, P.J. Jost, E. Heitzer, N. Dandachi, M. Balic, Integrating baseline ctDNA-derived tumor metrics enhances risk stratification in HR-positive/HER2-negative advanced breast cancer: a real-world multicenter cohort study from Austria, ESMO Open, https://doi.org/10.1016/j.esmoop.2026.106939

The post April 2026 Customer Publications appeared first on The Golden Helix Blog.

Long-Read Data

Long-read sequencing technologies like PacBio HiFi have transformed our ability to detect structural variants (SVs) with greater accuracy and resolution than ever before. Unlike short-read sequencing, which often struggles to span the repetitive or complex genomic regions where SVs are most common, long reads can reliably resolve insertions, deletions, inversions, duplications, and translocations that would otherwise go undetected. At Golden Helix, we’ve built comprehensive support for PacBio long-read SV analysis directly into VSWarehouse and VarSeq, covering everything from raw variant calling to clinical interpretation and report generation.

Ready-to-Use PacBio Workflows in VSWarehouse

VSWarehouse includes two purpose-built workflows for PacBio whole-genome sequencing data, both available for download from our Workflow Repository:

• PacBio Germline HiFi WGS Workflow: designed for constitutional variant analysis, this workflow uses Sawfish to call structural variants alongside small variants and CNVs.
• PacBio Somatic WGS Workflow: designed for tumor/cancer analysis, this workflow leverages a combination of Severus, Wakhan, and CNVKit to provide a comprehensive picture of somatic structural variation.

Both workflows are fully automated and culminate in the generation of a VarSeq project with structural variants, small variants, and CNVs already imported and ready for analysis.

Annotating and Filtering Structural Variants

Once a workflow completes, the resulting VarSeq project comes pre-configured with a set of annotations and filters tailored to structural variant interpretation. These templates are completely customizable, giving laboratories complete control over the annotation tracks, filter thresholds, and variant categories to match their specific clinical or research requirements.

The included annotations include gene information and effect predictions, allowing users to immediately distinguish between clinically distinct consequences, such as in-frame fusions or transcript ablations.

The screenshot below shows the filtered set of SVs called by Sawfish in our PacBio Germline HiFi workflow, illustrating how users can quickly identify a small list of potentially clinically relevant structural variants:

Clinical Interpretation in VSClinical

Once filtered, structural variants can be directly imported into VSClinical for interpretation and classification. VSClinical provides a rich set of tools specifically designed to support SV review, including variant summary information and breakpoint visualization.

The VSClinical visualization makes it clear that this chromosome 4 inversion disrupts the 5′ coding region of ANK2. The breakpoint occurs after exon 1, producing an ANK2::CAMK2D fusion transcript that is predicted to introduce a frameshift early in the ANK2 coding sequence.

VSClinical also integrates information from resource like ClinGen and OMIM including detailed references with links to PubMed, streamlining the process of gathering the supporting evidence required to draft variant interpretations.

Once structural variants have been interpreted and classified, they can be incorporated into a finalized clinical report using VSClinical’s customizable report templates. Reports are generated automatically based on the variants and interpretations captured during the review process, ensuring consistency and saving time.

Conclusion

As long-read sequencing continues to expand in both research and clinical settings, having a scalable and customizable analysis framework is essential. With VSWarehouse and VarSeq, laboratories can confidently leverage long read data to uncover clinically meaningful structural variants and bring clarity to even the most complex genomic rearrangements. Please reach out to us if you would like to learn more about support for long-read analysis in VarSeq and VSWarehouse.

The post Long-Read Part 1: Structural Variant Analysis appeared first on The Golden Helix Blog.

Every year on DNA Day, I am reminded that few scientific discoveries have had such a long and compounding arc of impact as the elucidation of the DNA double helix. What began as a fundamental insight into the structure of life has steadily evolved into one of the most practical tools in modern medicine. For me personally, DNA Day is not an abstract celebration of science. It is a reflection point on a professional journey that has been shaped by genomics long before it became fashionable, commercialized, or synonymous with next-generation sequencing. 

Early in my career, I studied genomics when the field was still largely theoretical operating on small data sets. This was well before NGS entered the mainstream conversation, before sequencing was fast, inexpensive, or widely accessible. Genomics then was about understanding structure, variation, and biological consequences, often with limited data and even more limited tooling. That early exposure shaped how I think about DNA to this day, not as raw sequence, but as information that only becomes valuable when it can be interpreted, contextualized, and trusted. That perspective has stayed with me throughout my career and ultimately led me to my role as CEO of Golden Helix. 

From Early Genomics to Real-World Impact 

When I first engaged with genomics, the promise was clear but the path to clinical impact was not. Sequencing was slow and expensive, and the gap between genetic data and medical decision-making was wide. What fascinated me even then was not the act of sequencing itself, but the question of interpretation. What does a variant mean? When does it matter? How confident can we be in acting on it? 

These questions predate NGS, and they remain central today. 

As genomics accelerated, especially with the arrival of high-throughput sequencing, the industry understandably focused on data generation. But data alone does not improve patient outcomes. Sound interpretation does, as do reliable workflows and responsible governance. This is where many early genomic efforts struggled, and where long-term value is still won or lost. 

DNA Day and the Reality of Clinical Genomics 

Today, DNA-driven medicine is no longer aspirational. It is operational across oncology, rare disease diagnostics, reproductive health, and pharmacogenomics. Yet despite the progress, we are still early in the lifecycle of true genomic medicine. 

One lesson that has become increasingly clear to me as a leader is that science moves faster than systems. We can generate extraordinary volumes of genomic data, but without robust interpretation frameworks, clinical workflows, and accountability, that data does not translate into better care. 

DNA Day is therefore as much about infrastructure as it is about discovery. 

At Golden Helix, our focus has always been on that foundational layer. Variant interpretation, reproducibility, auditability, and clinical defensibility are not afterthoughts. They are prerequisites for scaling genomics responsibly. These principles trace directly back to my early exposure to genomics, when careful interpretation mattered more than raw throughput. 

Scaling Trust Across Healthcare 

One of the defining challenges in genomic medicine is trust. Clinicians must trust that a reported interpretation is correct, current, and explainable. Laboratories must trust that their processes will stand up to regulatory scrutiny. Health systems must trust that genomic programs can scale without introducing risk or inconsistency. 

Trust does not come from novelty. It comes from rigor. 

Golden Helix’s long-standing engagement with academic partners and its history of NIH-funded research have played a critical role in shaping this mindset. Research funded in the public interest demands transparency, reproducibility, and methodological discipline. Those same values are essential when genomics moves into regulated clinical environments, where decisions affect real patients and real outcomes. 

A Long-Term View of Genomic Medicine 

From a leadership standpoint, genomics has reinforced the importance of playing the long game. Unlike many areas of technology, progress in this field is cumulative. Decisions made today about data models, standards, and interpretation logic will influence patient care for decades. 

DNA Day is a reminder that we are stewards of an enduring scientific legacy. The DNA molecule itself has not changed, but our responsibility in how we use it has grown. We are no longer simply discovering biology. We are operationalizing it. 

I often remind teams that our true challenge is not competition from other vendors, but inertia within healthcare systems. Changing how medicine is practiced requires confidence that genomic tools are not just powerful today, but sustainable tomorrow. 

Looking Ahead 

As we look forward, the role of DNA in medicine will only expand. Pharmacogenomics will increasingly guide prescribing decisions at scale. Whole-genome sequencing will move closer to first-line testing. Long-read sequencing and multi-omics will deepen biological insight while raising the bar for interpretation quality. 

Whether these advances fulfill their promise will depend less on technological breakthroughs and more on disciplined implementation. 

DNA Day, for me, is both a celebration and a checkpoint. It marks how far the field has come since my early days studying genomics, when the data was sparse but the questions were already profound. It also reinforces the responsibility we carry today to build systems that honor the science and serve patients reliably. At Golden Helix, that remains our focus. Turning DNA into durable clinical knowledge, and ensuring that this remarkable molecule continues to improve lives in practical, measurable ways. 

Happy DNA Day.

The post DNA Day: Why This Molecule Still Shapes My Journey and the Future of Medicine  appeared first on The Golden Helix Blog.

We’re grateful to everyone who tuned in for our April 15th webcast, “CI-SpliceAI for Splice Site Prediction and Variant Interpretation in VarSeq,” with presenter Nathan Fortier.

If you weren’t able to attend live, here’s a quick overview of what was covered: the ClinGen SVI Splicing Subgroup has released evidence-based recommendations for applying splice prediction scores within the ACMG/AMP variant classification framework, and commercial laboratories have long faced a licensing barrier with the gold-standard SpliceAI tool. This webcast walked through both of those topics and introduced CI-SpliceAI as a validated, commercially licensed alternative now integrated directly into VarSeq.

Webcast Highlights

A major focus of the webcast was the ClinGen SVI Splicing Subgroup’s recommendations, which provide the concrete thresholds and criteria mappings that the original 2015 ACMG/AMP guidelines left unspecified. The recommendations center on SpliceAI delta scores and address five criteria:

1. PVS1 is reserved exclusively for variants directly disrupting canonical ±1/±2 splice dinucleotides;
2. PP3 applies when the maximum SpliceAI delta score is 0.2 or above;
3. PS1 can be combined with PP3 when a previously classified variant shares the same predicted splicing impact; The delta position reported by SpliceAI (the predicted distance from a variant to the affected splice site) is what makes it possible to determine whether two variants share the same predicted impact.
4. BP4 applies when the maximum delta score is 0.1 or below;
5. BP7 can be combined with BP4 for synonymous and intronic variants.

The webcast also addressed a practical problem facing commercial clinical laboratories: SpliceAI’s pre-trained model weights and precomputed scores are restricted to non-commercial use. To determine whether an open-source alternative could serve as a clinical replacement, Golden Helix conducted an independent benchmarking study comparing SpliceAI, CI-SpliceAI, and OpenSpliceAI across three independent datasets.

The headline finding was that CI-SpliceAI achieves statistically equivalent performance to SpliceAI across all three benchmarks, with balanced Spearman correlation across all four delta score types and exact splice site match rates above 90%. OpenSpliceAI performed well on splice loss events but showed notably weaker correlation on splice gain events. Based on these results, CI-SpliceAI was identified as the preferred open-source alternative for clinical use.

Finally, the webcast introduced two new CI-SpliceAI annotation tracks available now in VarSeq for all users with a VSClinical license:

• The CI-SpliceAI Variant Track provides over 47 million precomputed scores natively computed for both GRCh37 and GRCh38, avoiding the liftover artifacts present in Illumina-provided SpliceAI scores and supplemented with ClinVar and other curated variants for comprehensive clinical coverage.
• The High Confidence Regions Track solves a critical interpretation problem by marking genomic positions where an absent variant can be confidently treated as scoring below the 0.1 threshold. BP4 can be applied when all of the following are true:
  • The variant is absent from the CI-SpliceAI track;
  • It overlaps a High Confidence Region;
  • It is a SNP, single-base insertion, or deletion of one to four base pairs.

A live demo walked through a hypothetical case of a patient with renal insufficiency, hypertension, and respiratory distress whose exome sequencing came back negative. CI-SpliceAI was used within VarSeq to identify a deeply intronic cryptic splice variant that would otherwise have been missed. Then, VSClinical was used to review delta scores, visualize the predicted cryptic acceptor site, and generate a final clinical report.

Looking ahead, full automation is coming later this year: the VarSeq auto-classifier will automatically evaluate CI-SpliceAI scores and apply all four evidence codes (PP3, PS1, BP4, and BP7) per ClinGen SVI recommendations, and VSClinical will gain a new splice site visualization tool that displays predicted cryptic splice site locations relative to existing exon structure.

Conclusion

If you missed the live session or would like to revisit any part of the discussion, the full webcast recording is available on demand here. For those interested in diving deeper into our benchmarking methodology and results, our preprint is available on bioRxiv here. If you’re interested in exploring how VarSeq and VSClinical can support splice variant interpretation in your clinical workflows, please contact our technical and sales experts to schedule an evaluation or demo.

The post Webcast Recap: CI-SpliceAI for Splice Site Prediction and Variant Interpretation in VarSeq appeared first on The Golden Helix Blog.

Many labs already use DRAGEN upstream, but want more control over what happens after variant calling is complete. VSWarehouse 3 connects those outputs to downstream tertiary workflows that labs can automate, configure, and adapt to their own interpretation and reporting process.

Using DRAGEN with VSWarehouse 3: A Better Path to Tertiary Analysis

Many labs already have secondary analysis in place. For Illumina users, that often means DRAGEN outputs stored in BaseSpace. In the earlier posts in this series, we focused on native upstream automation in VSWarehouse 3 (VSW3), including native support for WDL and built-in Sentieon workflows. Those examples highlight one side of the platform: when secondary and tertiary analysis are linked inside the same governed environment, it becomes easier to launch workflows, track inputs and outputs, and move directly into downstream interpretation, reducing turnaround time. While native execution is valuable, many labs want to keep their current upstream secondary analysis process and improve what happens after it finishes.

A better handoff from upstream analysis

For labs using DRAGEN, the approach is straightforward: secondary analysis runs upstream, outputs are stored in BaseSpace, and Warehouse automation pulls the needed files into the workspace so downstream interpretation can begin.

In day-to-day practice, the inefficient part of the workflow is often not alignment or variant calling itself. It is what happens immediately after: identifying the correct files, moving them into the right place, launching the next step, and making sure the interpretation workflow starts with the right inputs. When those transitions depend on too many manual steps, the process becomes harder to standardize and harder to scale. Clinical labs feel that burden in operational overhead and in the difficulty of maintaining a clear and consistent path from upstream processing into case review and reporting.

VSW3 helps by giving that transition more structure. Even when upstream analysis happens elsewhere, Warehouse can still provide a controlled environment for organizing files, launching downstream workflows, and preserving the context around each run. That makes it easier to connect external secondary results to interpretation and reporting without relying on ad hoc file handling or one-off local habits.

Integration with external secondary analysis pipelines

This is also where the broader VSW3 automation model becomes relevant. Tasks and workflows in Warehouse provide a framework for defining parameters, stages, dependencies, outputs, and run history in a way that is reviewable and repeatable. Not every lab needs to start by moving its entire upstream pipeline into Warehouse, but it matters that the platform can support both models: native automation where that makes sense, and integration where that is the better operational fit.

For clinical labs, that flexibility is important. Some teams want to keep the secondary pipeline they already trust while improving the path into case review, interpretation, and reporting. They are not looking for a disruptive rip-and-replace project. They are looking for a better bridge between sequencing outputs and the downstream work that matters most clinically. VSW3 supports that approach by giving labs a way to bring externally generated results into a more governed interpretation environment.

That point is also relevant for labs evaluating downstream platforms more broadly. Some environments are designed as tightly managed end-to-end systems. Others give labs more control over how workflows are configured, how automation is extended, and how tertiary analysis fits into their own process. Golden Helix has long emphasized the second model. For labs already using DRAGEN upstream, that can be an attractive combination: keep the upstream workflow you know, while gaining a more flexible downstream environment for interpretation and reporting.

Connecting secondary and tertiary analysis more effectively

The larger point of this series is that secondary and tertiary analysis work better when they are connected more intentionally. Sometimes that means native workflow execution in VSW3, as in the earlier WDL and Sentieon examples. Sometimes it means bringing externally generated outputs into Warehouse so the downstream process becomes more consistent, easier to manage, and less dependent on manual handoff.

For labs already using DRAGEN and BaseSpace, that is a practical entry point into VSWarehouse 3. You do not need to rebuild your upstream pipeline to improve the path into tertiary analysis. You need a downstream workflow that gives the handoff more structure, more consistency, and a clearer path into interpretation and reporting.

The post Warehouse Workflow Spotlight: Part 3 appeared first on The Golden Helix Blog.