A machine learning model to determine the accuracy of variant calls in capture-based next generation sequencing

Correlation between variant call accuracy and quality parameters in comprehensive cancer genomic profiling tests

Background

Comprehensive genomic profiling (CGP) tests have been widely utilized in clinical practice. In this test, the variant list automatically output from the data analysis pipeline often contains false-positive variants, although the correlation between the quality parameters and prevalence of false-positive variants remains unclear.

Methods

We analyzed 125 CGP tests performed in our laboratory. False-positive variants were manually detected via visual inspection. The quality parameters of both wet and dry processes were also analyzed.

Results

Among the 125 tests, 52 (41.6%) required more than one correction of the called variants, and 21 (16.8%) required multiple corrections. A significant correlation was detected between somatic false-positive variants and quality parameters in the wet (ΔΔCq, pre-capture library peak size, pre-capture library DNA amount, capture library peak size, and capture library concentration) and dry processes (total reads, mapping rates, duplication rates, mean depth, and depth coverage). Capture library concentration and mean depth were strong independent predictors of somatic false-positive variants.

Conclusions

We demonstrated a correlation between somatic false-positive variants and quality parameters in the CGP test. This study facilitates gaining a better understanding of CGP test quality management.

Artificial intelligence and database for NGS-based diagnosis in rare disease

Rare diseases (RDs) are rare complex genetic diseases affecting a conservative estimate of 300 million people worldwide. Recent Next-Generation Sequencing (NGS) studies are unraveling the underlying genetic heterogeneity of this group of diseases. NGS-based methods used in RDs studies have improved the diagnosis and management of RDs. Concomitantly, a suite of bioinformatics tools has been developed to sort through big data generated by NGS to understand RDs better. However, there are concerns regarding the lack of consistency among different methods, primarily linked to factors such as the lack of uniformity in input and output formats, the absence of a standardized measure for predictive accuracy, and the regularity of updates to the annotation database. Today, artificial intelligence (AI), particularly deep learning, is widely used in a variety of biological contexts, changing the healthcare system. AI has demonstrated promising capabilities in boosting variant calling precision, refining variant prediction, and enhancing the user-friendliness of electronic health record (EHR) systems in NGS-based diagnostics. This paper reviews the state of the art of AI in NGS-based genetics, and its future directions and challenges. It also compare several rare disease databases.

Improving the filtering of false positive single nucleotide variations by combining genomic features with quality metrics

MOTIVATION

Technical errors in sequencing or bioinformatics steps and difficulties in alignment at some genomic sites result in false positive (FP) variants. Filtering based on quality metrics is a common method for detecting FP variants, but setting thresholds to reduce FP rates may reduce the number of true positive variants by overlooking the more complex relationships between features. The goal of this study is to develop a machine learning-based model for identifying FPs that integrates quality metrics with genomic features and with the feature interpretability property to provide insights into model results.

RESULTS

We propose a random forest-based model that utilizes genomic features to improve identification of FPs. Further examination of the features shows that the newly introduced features have an important impact on the prediction of variants misclassified by VEF, GATK-CNN, and GARFIELD, recently introduced FP detection systems. We applied cost-sensitive training to avoid errors in misclassification of true variants and developed a model that provides a robust mechanism against misclassification of true variants while increasing the prediction rate of FP variants. This model can be easily re-trained when factors such as experimental protocols might alter the FP distribution. In addition, it has an interpretability mechanism that allows users to understand the impact of features on the model's predictions.

AVAILABILITY AND IMPLEMENTATION

The software implementation can be found at https://github.com/ideateknoloji/FPDetect.

Confirmation of Insertion, Deletion, and Deletion-Insertion Variants Detected by Next-Generation Sequencing

BACKGROUND

Despite clinically demonstrated accuracy in next generation sequencing (NGS) data, many clinical laboratories continue to confirm variants with Sanger sequencing, which increases cost of testing and turnaround time. Several studies have assessed the accuracy of NGS in detecting single nucleotide variants; however, less has been reported about insertion, deletion, and deletion-insertion variants (indels).

METHODS

We performed a retrospective analysis from 2015-2022 of indel results from a subset of NGS targeted gene panel tests offered through the Mayo Clinic Genomics Laboratories. We compared results from NGS and Sanger sequencing of indels observed in clinical runs and during the intra-assay validation of the tests.

RESULTS

Results demonstrated 100% concordance between NGS and Sanger sequencing for over 490 indels (217 unique), ranging in size from 1 to 68 basepairs (bp). The majority of indels were deletions (77%) and 1 to 5 bp in length (90%). Variant frequencies ranged from 11.4% to 67.4% and 85.1% to 100% for heterozygous and homozygous variants, respectively, with a median depth of coverage of 2562×. A subset of indels (7%) were located in complex regions of the genome, and these were accurately detected by NGS. We also demonstrated 100% reproducibility of indel detection (n = 179) during intra-assay validation.

CONCLUSIONS

Together this data demonstrates that reportable indel variants up to 68 bp can be accurately assessed using NGS, even when they occur in complex regions. Depending on the complexity of the region or variant, Sanger sequence confirmation of indels is usually not necessary if the variants meet appropriate coverage and allele frequency thresholds.

Recommendations for Next-Generation Sequencing Germline Variant Confirmation: A Joint Report of the Association for Molecular Pathology and National Society of Genetic Counselors

Clinical laboratory implementation of next-generation sequencing (NGS)-based constitutional genetic testing has been rapid and widespread. In the absence of widely adopted comprehensive guidance, there remains substantial variability among laboratories in the practice of NGS. One issue of sustained discussion in the field is whether and to what extent orthogonal confirmation of genetic variants identified by NGS is necessary or helpful. The Association for Molecular Pathology Clinical Practice Committee convened the NGS Germline Variant Confirmation Working Group to assess current evidence regarding orthogonal confirmation and to establish recommendations for standardizing orthogonal confirmation practices to support quality patient care. On the basis of the results of a survey of the literature, a survey of laboratory practices, and subject expert matter consensus, eight recommendations are presented, providing a common framework for clinical laboratory professionals to develop or refine individualized laboratory policies and procedures regarding orthogonal confirmation of germline variants detected by NGS.

Satellite DNAs and human sex chromosome variation

Satellite DNAs are present on every chromosome in the cell and are typically enriched in repetitive, heterochromatic parts of the human genome. Sex chromosomes represent a unique genomic and epigenetic context. In this review, we first report what is known about satellite DNA biology on human X and Y chromosomes, including repeat content and organization, as well as satellite variation in typical euploid individuals. Then, we review sex chromosome aneuploidies that are among the most common types of aneuploidies in the general population, and are better tolerated than autosomal aneuploidies. This is demonstrated also by the fact that aging is associated with the loss of the X, and especially the Y chromosome. In addition, supernumerary sex chromosomes enable us to study general processes in a cell, such as analyzing heterochromatin dosage (i.e. additional Barr bodies and long heterochromatin arrays on Yq) and their downstream consequences. Finally, genomic and epigenetic organization and regulation of satellite DNA could influence chromosome stability and lead to aneuploidy. In this review, we argue that the complete annotation of satellite DNA on sex chromosomes in human, and especially in centromeric regions, will aid in explaining the prevalence and the consequences of sex chromosome aneuploidies.

Comprehensive Validation of Diagnostic Next-Generation Sequencing Panels for Acute Myeloid Leukemia Patients

Next-generation sequencing has greatly advanced the molecular diagnostics of malignant hematological diseases and provides useful information for clinical decision making. Studies have shown that certain mutations are associated with prognosis and have a direct impact on treatment of affected patients. Therefore, reliable detection of pathogenic variants is critically important. In this study, we aimed to compare four sequencing panels with different characteristics, from number of genes covered to technical aspects of library preparation and data analysis workflows, to find the panel with the best clinical utility for myeloid neoplasms with a special focus on acute myeloid leukemia. Using the Acrometrix Oncology Hotspot Control DNA and DNA from acute myeloid leukemia patients, we evaluated panel performance in terms of coverage, precision, recall, and reproducibility and tested different bioinformatics tools that can be used for the evaluation of any next-generation sequencing panel. Taken together, our results support the reliability of the Acrometrix Oncology Hotspot Control to validate and compare sequencing panels for hematological diseases and show which panel-software combination (platform) has the best performance.

Long-read nanopore sequencing enables accurate confirmation of a recurrent PMS2 insertion-deletion variant located in a region of complex genomic architecture

Targeted next generation sequencing (NGS) is the predominant methodology for the molecular genetic diagnosis of inherited conditions. In many laboratories, NGS-identified variants are routinely validated using a different method, to minimize the risk of a false-positive diagnosis. This can be particularly important when pathogenic variants are located in complex genomic regions. In this situation, new long-read sequencing technologies have potential advantages over existing alternatives. However, practical examples of their utility for diagnostic purposes remain scant. Here, we report the use of nanopore sequencing to validate a PMS2 mutation refractory to conventional methods. In a patient who presented with colorectal cancer and loss of PMS2 immunostaining, short-read NGS of Lynch syndrome-associated genes identified the recurrent PMS2 insertion-deletion variant, c.736_741delinsTGTGTGTGAAG (p.Pro246Cysfs*3). Confirmation of this variant using bidirectional Sanger sequencing was impeded by an upstream intron 6 poly(T) tract. Using a locus-specific amplicon template, we undertook nanopore long-read sequencing in order to assess the diagnostic accuracy of this platform. Pairwise comparison between a curated benchmark allele (derived from short-read NGS and unidirectional Sanger sequencing) and the consensus nanopore dataset revealed 100% sequence identity. Our experience provides insight into the robustness and ease of deployment of "third-generation" sequencing for accurate characterisation of pathogenic variants.

Prevalence of Inherited Mutations in Breast Cancer Predisposition Genes among Women in Uganda and Cameroon

BACKGROUND

Sub-Saharan Africa (SSA) has a high proportion of premenopausal hormone receptor negative breast cancer. Previous studies reported a strikingly high prevalence of germline mutations in BRCA1 and BRCA2 among Nigerian patients with breast cancer. It is unknown if this exists in other SSA countries.

METHODS

Breast cancer cases, unselected for age at diagnosis and family history, were recruited from tertiary hospitals in Kampala, Uganda and Yaoundé, Cameroon. Controls were women without breast cancer recruited from the same hospitals and age-matched to cases. A multigene sequencing panel was used to test for germline mutations.

RESULTS

There were 196 cases and 185 controls with a mean age of 46.2 and 46.6 years for cases and controls, respectively. Among cases, 15.8% carried a pathogenic or likely pathogenic mutation in a breast cancer susceptibility gene: 5.6% in BRCA1, 5.6% in BRCA2, 1.5% in ATM, 1% in PALB2, 0.5% in BARD1, 0.5% in CDH1, and 0.5% in TP53. Among controls, 1.6% carried a mutation in one of these genes. Cases were 11-fold more likely to carry a mutation compared with controls (OR = 11.34; 95% confidence interval, 3.44-59.06; P < 0.001). The mean age of cases with BRCA1 mutations was 38.3 years compared with 46.7 years among other cases without such mutations (P = 0.03).

CONCLUSIONS

Our findings replicate the earlier report of a high proportion of mutations in BRCA1/2 among patients with symptomatic breast cancer in SSA.

IMPACT

Given the high burden of inherited breast cancer in SSA countries, genetic risk assessment could be integrated into national cancer control plans.

Methods for Identifying Patients with Tropomyosin Receptor Kinase (TRK) Fusion Cancer

NTRK gene fusions affecting the tropomyosin receptor kinase (TRK) protein family have been found to be oncogenic drivers in a broad range of cancers. Small molecule inhibitors targeting TRK activity, such as the recently Food and Drug Administration-approved agent larotrectinib (Vitrakvi®), have shown promising efficacy and safety data in the treatment of patients with TRK fusion cancers. NTRK gene fusions can be detected using several different approaches, including fluorescent in situ hybridization, reverse transcription polymerase chain reaction, immunohistochemistry, next-generation sequencing, and ribonucleic acid-based multiplexed assays. Identifying patients with cancers that harbor NTRK gene fusions will optimize treatment outcomes by providing targeted precision therapy.

Privacy and ethical challenges in next-generation sequencing

INTRODUCTION

Next-generation sequencing (NGS) is expected to revolutionize health care. NGS allows for sequencing of the whole genome more cheaply and quickly than previous techniques. NGS offers opportunities to advance medical diagnostics and treatments, but also raises complicated ethical questions that need to be addressed.

AREAS CONSIDERED

This article draws from the literature on research and clinical ethics, as well as next-generation sequencing, in order to provide an overview of the ethical challenges involved in next-generation sequencing. This article includes a discussion of the ethics of NGS in research and clinical contexts.

EXPERT OPINION

The use of NGS in clinical and research contexts has features that pose challenges for traditional ethical frameworks for protecting research participants and patients. NGS generates massive amounts of data and results that vary in terms of known clinical relevance. It is important to determine appropriate processes for protecting, managing and communicating the data. The use of machine learning for sequencing and interpretation of genomic data also raises concerns in terms of the potential for bias and potential implications for fiduciary obligations. NGS poses particular challenges in three main ethical areas: privacy, informed consent, and return of results.

A Rigorous Interlaboratory Examination of the Need to Confirm Next-Generation Sequencing-Detected Variants with an Orthogonal Method in Clinical Genetic Testing

Orthogonal confirmation of next-generation sequencing (NGS)-detected germline variants is standard practice, although published studies have suggested that confirmation of the highest-quality calls may not always be necessary. The key question is how laboratories can establish criteria that consistently identify those NGS calls that require confirmation. Most prior studies addressing this question have had limitations: they have been generally of small scale, omitted statistical justification, and explored limited aspects of underlying data. The rigorous definition of criteria that separate high-accuracy NGS calls from those that may or may not be true remains a crucial issue. We analyzed five reference samples and over 80,000 patient specimens from two laboratories. Quality metrics were examined for approximately 200,000 NGS calls with orthogonal data, including 1662 false positives. A classification algorithm used these data to identify a battery of criteria that flag 100% of false positives as requiring confirmation (CI lower bound, 98.5% to 99.8%, depending on variant type) while minimizing the number of flagged true positives. These criteria identify false positives that the previously published criteria miss. Sampling analysis showed that smaller data sets resulted in less effective criteria. Our methodology for determining test- and laboratory-specific criteria can be generalized into a practical approach that can be used by laboratories to reduce the cost and time burdens of confirmation without affecting clinical accuracy.

Using Big Data and Predictive Analytics to Determine Patient Risk in Oncology

Big data and predictive analytics have immense potential to improve risk stratification, particularly in data-rich fields like oncology. This article reviews the literature published on use cases and challenges in applying predictive analytics to improve risk stratification in oncology. We characterized evidence-based use cases of predictive analytics in oncology into three distinct fields: (1) population health management, (2) radiomics, and (3) pathology. We then highlight promising future use cases of predictive analytics in clinical decision support and genomic risk stratification. We conclude by describing challenges in the future applications of big data in oncology, namely (1) difficulties in acquisition of comprehensive data and endpoints, (2) the lack of prospective validation of predictive tools, and (3) the risk of automating bias in observational datasets. If such challenges can be overcome, computational techniques for clinical risk stratification will in short order improve clinical risk stratification for patients with cancer.