Table 3.

Considerations for Molecular Genetic Testing

A. Types of molecular genetic variants *
Those affecting exonic (coding) sequence.
Missense	Single base variant that changes an amino acid
Nonsense	Single base variant that changes an amino acid to a stop codon
Insertion/deletion (indel)	Usually one or a few nucleotides inserted or deleted. Unless the indel is in a multiple of three, a frameshift occurs that garbles the usual amino acid sequence. This usually results in an eventual stop codon.
Those affecting intronic or splice site sequences
Splice site	Affects exon splicing; one or more exons may be skipped
Intronic	By definition intronic sequencing is non-coding. While intronic variation is more common than coding sequence, it has been infrequently associated with disease.

B. Testing categories of sequence variations relevant to a phenotype of interest (90)

Sequence variation is previously reported and is a recognized cause of the disorder Sequence variation is previously unreported and is of the type which is expected to cause the disorder Sequence variation is previously unreported and is of the type which may or may not be causative of the disorder (also commonly referred to as a variant of unknown significance – VUS) Sequence variation is previously unreported and is probably not causative of disease Sequence variation is previously reported and is a recognized neutral variant Sequence variation is not known or expected to be causative of disase, but is found to be associated with a clinical presentation

C. Criteria used to assess the relevance of a genetic variant for a phenotype of interest
Property	Comment
Prior molecular genetic classification, if available	This may be definitive for variants previously established as disease-causing.
Type of variant (see Section A in this table).	A synonymous variant only in unusual circumstances is considered relevant for disease (e.g., a variant that opens a cryptic splice acceptor site).
Weight of evidence, in the gene in question, that rare nonsynonymous variants cause DCM.	**see comment below.
This is especially relevant for a novel gene under consideration in a discovery study. Disruption of a functional protein in the tissue of interest could lead to plausible pathophysiology.	Examples of established genes include those encoding proteins of the contractile apparatus (see Table 1). For discovery studies, evidence of cardiac expression or the presence of the protein product in cardiac tissue may aid in assessing relevance.
Rarity in the population	Many Mendelian variants may be ‘private’ or unique to a proband or family
Variant segregates with the DCM phenotype, ideally in one or more large families; lacking large families, the variant segregates with DCM in multiple smaller families, or is observed in multiple sporadic DCM cases.	In genetic DCM (and other multi-locus Mendelian diseases), many variants are ‘private’ so that multiple probands or families with any one specific variant are uncommon.
Functional data derived from the variant: cellular or animal models that recapitulate the disease phenotype	All model systems have inherent limitations and seldom provide definitive studies; however, such functional data increases the certainty that the variant under study is relevant for phenotype of interest.

^*These variants do not account for copy number variants (CNV's; also termed structural variants), which are insertions, deletions, duplications or inversions of larger portions of DNA. CNVs range widely in size; from very small (fewer than a hundred nucleotides) to very large (many megabases), and all sizes in between. They may affect both coding and non-coding DNA. Structural variants are not detected by usual sequencing approaches. Systematic evaluation of structural variants has not been undertaken in DCM, and hence their relevance for DCM has not been established.

^**Some genes (e.g., LMNA, MYH7, TNNT2, see Table 1) have abundant evidence that point mutations can cause DCM. Nevertheless, because of the marked allelic heterogeneity in DCM genes, it is uncommon for any one specific variant to be found in multiple unrelated probands, even in these genes. Whether any of these novel nonsynonymous rare variants can be considered disease-causing by usual molecular genetic diagnostic standards is an open question. Further, because most of the DCM genes (Table 1) have had only a few reported pathologic variants, newly identified rare variants in such genes with less prior DCM sequencing data available are commonly reported as variants of unknown significance (VUS; this table, Section B).