现在的位置: 首页研究点评, 进展交流>正文
[JAMA观察]:RNA测序作为诊断手段
2023年01月24日 研究点评, 进展交流 [JAMA观察]:RNA测序作为诊断手段已关闭评论

JAMA Insights 

Genomics and Precision Health

December 16, 2022

RNA Sequencing as a Diagnostic Tool

Shamika Ketkar, Lindsay C. Burrage, Brendan Lee

JAMA. Published online December 16, 2022. doi:10.1001/jama.2022.22843

RNA sequencing (RNA-seq) is a new tool in the genetic diagnostic laboratory made possible by the advent of low-cost, high-throughput, next-generation sequencing technology. Historically, RNA studies were limited to gene expression using microarray technology for the detection of gene rearrangements and quantification of gene expression for predefined genes. RNA-seq allows detection of qualitative and quantitative changes in RNA expression across the genome in clinical samples and is increasingly being used as an adjunct to diagnostic exome sequencing and whole-genome sequencing.

How It Works

RNA-seq works by sequencing every RNA molecule in a sample and measuring the expression of genes by counting the number of times its transcripts have been sequenced. The resulting information can be used to identify aberrant transcript events (genes expressed at aberrant levels), aberrantly spliced genes (where RNA transcripts are inaccurately cleaved to remove introns to become mature mRNA), and expressed rare variants where only 1 of the 2 copies of a gene is active while the other is silent.

A typical RNA-seq workflow begins with tissue sample selection followed by complementary DNA synthesis of total RNA by reverse transcription (Figure). Complementary DNA is then fragmented and sequencing adaptors are added to each end of these fragments, which are subjected to massively parallel sequencing to obtain sequence reads from one or both ends (single or paired-end sequencing, respectively). After sequencing, the resulting reads are assembled using either a reference-guided approach, ie, based on a template DNA blueprint, or de novo, ie, reference free using the information contained in the reads alone.

Several bioinformatic methods are available to analyze transcript abundance, fusion events (when 2 genes are rearranged in cell to create a novel mRNA), or splicing activity.1

Important Considerations

Molecular diagnostic RNA-seq has, to date, largely been conducted in the research arena. However, commercially available RNA-seq is increasingly available as an adjunct to DNA sequencing for cancer and rare disease clinical diagnostic purposes.

One important consideration for RNA-seq analysis is the selection of an appropriate tissue type for study. Disease-relevant tissue serves as the best sample for RNA-seq because gene expression and splicing patterns are unique to each tissue and cell type. For diagnostic purposes, RNA-seq is often performed on clinically accessible tissues such as whole blood, Epstein-Barr virus–transformed lymphocytes, skeletal muscle, or skin-derived fibroblasts from skin biopsies.2 Thus, aberrations in genes that are expressed only in nonclinically accessible tissues, such as brain or heart, may remain undetected.

Aicher et al3 quantified RNA splicing in 801 RNA-seq samples from 53 adult tissues and 3 fetal tissues (cerebellum, cortex, and heart) from freely available databases to determine whether clinically accessible tissues serve as a proxy for nonclinically accessible tissues for RNA splicing. In this study, the clinically accessible tissues inadequately represented 6.3% of genes with splicing events in nonclinically accessible tissues, whereas 40.2% of these genes were inadequately represented by at least 1 clinically accessible tissue.3 They further evaluated how inadequately represented splicing was related to low gene expression. Of the genes with splicing events in nonclinically accessible tissues, a majority (52.1%) had low expression in clinically accessible tissues (transcripts per million <1) but in contrast 5.8% of genes were found to be inadequately expressed despite databases indicating adequate expression (transcripts per million >10). Overall, skin-derived fibroblasts most adequately reflected gene splicing patterns in nonclinically accessible tissues compared with blood.3 Several studies confirmed that although blood is regarded as an easily accessible tissue, it covers only about 50% of genes listed in the Online Mendelian Inheritance in Man database compared with about 70% in fibroblasts. In addition, some pathogenic splicing variants detected in fibroblasts are not detected in blood or muscle.2,4

Another important consideration when using RNA analysis is whether aberrant splicing impacts protein function. Aberrant splicing could lead to loss of normal transcript but in some cases, multiple splice products, including some residual normal transcript, may be produced. Likewise, abnormal splicing patterns that are in frame may have little impact on protein function. Thus, an assay of protein function may still be necessary to interpret pathogenicity of variants of uncertain significance.

Clinicians should also recognize that at this time there are no widely accepted reference standards for assessing the analytic and clinical validity or clinical utility of RNA-seq assays used for clinical applications.

Value

The clinical utility of RNA-seq for evaluation of DNA variants of uncertain significance for pathogenicity has been demonstrated in precision oncology and the care of individuals with rare mendelian disorders. In undiagnosed patients, RNA-seq has provided evidence for the pathogenicity of variants that were previously missed or misinterpreted by exome sequencing or whole-genome sequencing, leading to a 7.5% to 36% improvement in the molecular diagnostic rate.2,4 The cost of goods for RNA-seq can vary depending on whether cell culture is required (ie, for fibroblasts) vs direct analysis of whole blood. In general, downstream reagent and sequencing costs are similar to exome sequencing, although clinical interpretation cost is a significant variable. As with exome sequencing, insurer coverage is widely variable and less available with RNA-seq outside of targeted cancer genetic testing.

Evidence Base

There are no evidence-based clinical guidelines that incorporate RNA-seq in rare disease or cancer diagnostics. The clinical diagnostic laboratories offering RNA testing primarily use it for determining the functional consequences of DNA variants of uncertain significance, with widest application in cancer genetics. Further clarification of variants of uncertain significance as disease-causing or benign may affect clinical management recommendations. RNA-seq has been used successfully to genetically diagnose patients with rare mendelian disorders mostly with inconclusive exome sequencing and whole-genome sequencing results in the context of research studies. Utilizing data from the Undiagnosed Diseases Network, Lee et al5 found an 18% molecular diagnostic yield in a heterogeneous cohort, and Murdock et al4 reported 17% yield (14/83) in exome sequencing–negative cases encompassing a variety of phenotypes. Most recently, Yépez et al2 unraveled the molecular etiology in 16% (33/205 fibroblast RNA-seq) of the cases with suspected mitochondrial disease. RNA-seq–first or –only approaches may be used to effectively direct exome sequencing or whole-genome sequencing analysis to possibly pathogenic genetic variations in the absence of a priori candidates.

Conclusions

RNA sequencing can improve molecular diagnostic rates achieved by diagnostic exome sequencing or whole-genome sequencing alone. RNA-seq may successfully detect pathogenic variants missed by limitations in the interpretation of DNA variants of uncertain significance. Implementation of RNA-seq in clinical diagnostics requires establishment and standardization of methods for assessing analytical and clinical validity as well as clinical utility.

抱歉!评论已关闭.

×
腾讯微博