As next-generation sequencing (NGS) becomes more embedded in clinical diagnostics and research workflows, the ability to confidently interpret and report on genomic findings is more critical than ever. In July 2025, Golden Helix’s software suite continued to play a central role in variant annotation and classification, with publications spanning diverse clinical contexts, from population-wide secondary findings to rare disease diagnostics and idiopathic conditions. These peer-reviewed studies underscore the growing global reach of NGS, demonstrating the value of how our software supports high-quality, guideline-driven interpretations and bridges the gap between data and clinical insight.
Certain vs. uncertain actionable secondary findings in a cohort of 500 Lebanese participants: What to report to the patient?
Advances in next-generation sequencing enabled its integration into genetic diagnosis and have led to the uncovering of secondary findings. In this paper, we analyzed 500 Lebanese participants for pathogenic and likely-pathogenic variants in 81 recommended genes listed by the American College of Medical Genetics (ACMG). In this retrospective study, 500 individuals seeking genetic diagnosis through Exome Sequencing were included. Variants were analyzed and their pathogenicity assessed based on ACMG/AMP criteria and ClinVar. Secondary findings were identified in 16.8% of cases based on ACMG/AMP criteria, which decreased to 6% when relying on ClinVar. Dominant cardiovascular disease variants were predominant, constituting 6.6% based on ACMG/AMP assessments and 2% according to ClinVar. Additionally, using ACMG/AMP criteria, dominant oncogenic variants were identified in 4.2% of individuals, while recessive pathogenic variants were found in 4.8%. In contrast, ClinVar-based analysis reported these variants in 1% and 2.6% of the cohort, respectively. The high discordance between ACMG/AMP and ClinVar classifications (16.8% vs. 6%) underscores ethical dilemmas in deciding which criteria to prioritize for patient disclosure. Indeed, the absence of ACMG-classified pathogenic or likely pathogenic variants in ClinVar complicates reporting due to a lack of evidence linking them to disease in other individuals. Finally, the significant discrepancy between ACMG/AMP and ClinVar classifications emphasizes the urgent need to harmonize variant databases and update ClinVar entries, particularly for understudied populations such as the Lebanese cohort.
“The pathogenicity of the filtered variants was then examined based on the guidelines for the interpretation of sequence variants established by the ACMG and the Association for Molecular Pathology (ACMG/AMP) [4] as well as on ClinVar database [14], and taking into consideration ClinGen Variant Curation Expert Panels (VCEPs). ACMG/AMP classification, was initially attributed to each variant, using several tools (such as InterVar, VSClinical, VarSome, Franklin); then confirmed by a manual verification to ensure accuracy and reliability. This enabled the identification of 63 P/LP variants in 84 individuals (Table 1 and S1 Table), while ClinVar identified 20 P/LP variants in 30 participants (Table 1). As expected, a statistically significant association between the two classification systems was observed (data not shown). Indeed, all P and LP variants present in ClinVar were also classified as P or LP by the ACMG/AMP guidelines (Table 1).”
Sequencing validates deep learning models for EHR-based detection of Noonan syndrome in pediatric patients
Despite advanced diagnostic tools, early detection of rare genetic conditions like Noonan syndrome (NS) remains challenging. We evaluated a deep learning model’s real-world performance in identifying potential NS cases using electronic health record (EHR) data, validated through genetic sequencing and clinical assessment. The model analyzed 92,428 patients, identifying 171 high-risk individuals (score > 0.8) who underwent comprehensive review. Among these, 86 had prior genetic diagnoses, including three NS cases diagnosed during the study period. Genetic sequencing of remaining patients identified two additional NS cases with pathogenic variants. The model achieved 2.92% precision and 99.82% specificity. While precision was lower than prior validation (33.3%), this reflected expected differences in disease prevalence rather than model degradation. NS-associated phenotypes were enriched among high-risk patients, and trajectory analysis showed potential for earlier identification, highlighting both promise and limitations of EHR-based computational screening tools.
“Variant annotation used an in-house R pipeline incorporating gnomAD23 for population minor allele frequencies, ClinVar24 for pathogenicity classifications, protein consequences (synonymous, nonsynonymous, nonsense, frameshift, and splice site variants), and MetaSVM25 for deleteriousness prediction based on data from ANNOVAR26. Variants retained met the following criteria: rare frequency (MAF < 0.1%), location in NS/congenital heart defect-related genes27 (Supplementary Table 1), moderate or high predicted consequences, adequate coverage (>8x), high genotype quality (≥99), not classified as “Benign” in ClinVar, and predicted deleterious by MetaSVM. Phasing analysis using TinkerHap28 was performed when multiple variants were identified in the same gene to confirm compound heterozygosity. A clinical geneticist (KNW) manually reviewed filtered variants according to ACMG guidelines29 for NS diagnosis confirmation. This analysis was complemented using VarSeq™ (Golden Helix, Inc., Bozeman, MT, www.goldenhelix.com) software.”
Two Cases Diagnosed with Idiopathic Root Resorption and Low Serum Vitamin D Raise New Questions on Aetiology
Aims: In this report, two cases, A and B, with idiopathic resorptions are presented. In both cases the hypothesis was that the idiopathic resorption processes had a general medical cause, presumably an inborn calcification deficit. The aim was to evaluate this hypothesis.
Presentation of Cases: Case A. Healthy Caucasian male, born 1999, with no anamnestic information on diseases or medications, was treated with orthodontic fixed appliances for agenesis of a mandibular incisor, lack of space in the maxilla for cuspid eruption and bilateral open bite. A sister had minor resorption defects after orthodontic treatment. What is extraordinary in case A, and seemingly not described before, is the aggressive resorption occurring in the retention period and in the 4-year post retention period.
Case B. Caucasian male, born in a pre-term delivery in 2003 with an anamnestic information on late development and ADHD. Case B has never received orthodontic treatment. Both patients underwent a serum test and case A was also offered a genetic test.
Findings: Cases A and B both had low values of vitamin D. In addition, case B had low value of alkaline phosphatase (ALP). Case A was genetically negative for hypophosphasia (HPP).
By retrospection, both cases revealed severe resorption in the primary dentition before onset of orthodontic treatment.
Conclusion: It was concluded that the resorptions observed in the permanent teeth in case A was not a consequence of the orthodontic treatment. It is suggested that case A could have osteomalasia, while case B may have HPP.
Limitations: This study represents a new approach in revealing the aetiology behind severe idiopathic root resorption. Further collaboration with medical specialists is need for improving the indications and the limitations for the serological methods.
“Blood for genetic screening was received from Case A early 2020. Due to the Covid-19 pandemic the University Laboratory was more or less closed until Summer 2020. Whole Genome Exome sequencing (WES) was performed at a commercial laboratory (Dante labs). Resulting fastq files were aligned to a human reference genome (GRCh37) and variants were called using GATK pipeline following the Best Practices workflow [9]. Filtering and prioritization of variants were done using VarSeq [10].”
Whether it’s refining variant classification across underrepresented populations, validating AI models for early genetic diagnoses, or uncovering new questions in dental genetics, these use cases from July 2025 reinforce the versatility and clinical value of the Golden Helix platform. VarSeq and VSClinical continue to serve as trusted solutions for labs worldwide, enabling scalable, reproducible, and transparent analyses. As the genomics industry advances, we’re proud to support customers in producing high-confidence results that directly impact patient care.