Molecular diagnostic strategies

A correct molecular diagnosis is essential for patients with genetic myopathies as identification of the causative mutation(s) facilitates informative genetic counseling, proper patient care, and management, as well as the administration of appropriate therapeutic interventions. Several diagnostic strategies exist for identifying pathogenic or putative pathogenic variants in patients with inherited myopathies.

These involve knowing the mode of inheritance in the pedigree, availability of appropriate muscle tissue, and family DNA material.

2.2.3.1 DNA sequencing

In an unsolved myopathy family, DNA is collected from patients and available and informative family members. The causative genes can be identified in cohorts with established genetic diagnostic methodologies for most of the currently known myopathies. However, most patients with myopathies referred for genetic testing remain without a final diagnosis due to yet unknown molecular mechanisms (Savarese et al., 2016).

Sequencing the coding regions of DNA or exons is referred to as Exome Sequencing (ES), previously Whole exome sequencing (WES). Several different capture-based methods are available to enrich, amplify and sequence fragmented DNA.

These methods utilize massive parallel sequencing (MPS), a type of high throughput sequencing (HTS). However, analyzing ES data requires heavy computational infrastructure and knowledge of sequencing data analysis. In a targeted gene panel sequencing, known causative genes and candidate genes

with similar biochemical properties are included. Several diagnostic laboratories have custom-targeted gene panels that aid in the rapid molecular diagnosis of rare disease patients.

Examples of these are the MYOcap (Evila et al., 2016) and the Motorplex (Savarese et al., 2014) gene panels. In addition to these, several commercial diagnostic panels are nowadays also available from Illumina, Agilent, and Roche, and other private diagnostic services.

Alternatively, MPS methods are also used to sequence the genomic regions of DNA, including the coding and the non-coding part, referred to as Genome Sequencing (GS), previously Whole Genome sequencing (WGS). In recent years ES and GS strategies are increasingly becoming standard diagnostic approaches for Mendelian diseases (Xue et al., 2015). However, some technical challenges remain. Non-uniform coverage issues of complex sequence regions, segmental duplications, and satellite regions or ‘holes’ have been identified in all capture-based methods (Meienberg et al., 2015; Barbitoff et al., 2020) suggesting that neither method yet sequence all the targeted regions, entire exome or genome with uniform and adequate depth.

2.2.3.2 RNA sequencing

Despite the success of DNA sequencing methods based on MPS, the current diagnostic rate in myopathies is low, and additional sequencing strategies like transcriptome sequencing can add more levels of information. High throughput sequencing of tissue-specific mRNA (RNA-seq) from skeletal muscle biopsies or RNA obtained from fibroblasts of the patients reflect the disease specificity and can aid in the molecular diagnosis of unsolved myopathies. Recent studies have shown an increase in the rate of detecting novel pathogenic mutations in neuromuscular disorders via RNA-seq (Cummings et al., 2017; Gonorazky et al., 2019). However, RNA-seq is not a standalone diagnostic tool, and the framework for guidelines and interpretations of variant

findings from RNA-seq for diagnostic uses is still in progress.

RNA-seq can also improve the interpretation of variants identified on DNA and provide information about possible allele-specific changes or splicing aberrations (Wang et al., 2009; Byron et al., 2016). Sequencing data from RNA have traditionally been used for differential expression profiling.

However, with the advancements in systems and network interaction studies, RNA-seq can be used to understand intramolecular and intermolecular RNA-RNA interactions (RRIs) or interactions with protein, thus providing valuable information on different biological processes in the interactome (Stark et al., 2019).

2.2.3.3 Variant pathogenicity

A significant challenge in analyzing sequencing data is to determine the pathogenicity of the variants. The American College of Medical Genetics and Genomics (ACMG), along with the Association of Molecular Pathology (AMP), has issued a set of guidelines for interpretation of variants in Mendelian diseases (Richards et al., 2015). These guidelines recommend the use of specific terms -"pathogenic," "likely pathogenic,"

"uncertain significance," "likely benign," and "benign"-to describe variants identified from MPS data. These guidelines suggest classifying the variants into these five categories based on different levels of evidence, e.g., minor allele frequency (MAF) in population data, in silico predictions, functional data, and segregation data.

Estimation of a variant frequency threshold can be made based on the mode of inheritance in the family and population frequency of the variant. Additionally, the segregation of variants with the phenotype in the family increases the value of the information provided for determining the variant pathogenicity. Consequences of the genetic variant on protein level are estimated with in silico prediction tools like CADD, SIFT, PolyPhen, and MutationTaster, but should just be

considered preliminary assumptions. Functional studies in cell or animal models can be highly informative in asserting the damaging role of the genetic variant. Strong evidence of pathogenicity of a previously not reported variant requires a very low MAF in control population or public databases along with segregation data (or a confirmed de novo) and well-established functional studies showing the deleterious effects of the variants. On the other hand, strong evidence of the benign effect of the variant is considered when the MAF is too high in the population, the variant fails to segregate with the phenotype, and functional studies show no harmful effect of the variant.

Approximately 40% of variants observed in massively parallel sequencing data can be categorized as variants of uncertain or unknown significance (VUS) (Federici & Soddu, 2020). The interpretation of VUS, in both clinical and research settings, remains challenging for all Mendelian and complex diseases. The use of additional ‘omics’ data like transcriptomics, proteomics or metabolomics, functional studies, and network-based gene association studies can significantly improve the molecular understanding of most of these variants. Additionally, the use of variant pathogenicity guidelines (Richards et al., 2015) is a must in determining the pathogenicity score of a particular variant concerning the phenotype and the families being studies.