Sequence Variants (SNV, indel, duplication)

Sequence variants are small-scale alterations of the DNA sequence detectable by short-read next-generation sequencing (NGS) and Sanger confirmation. They span single base substitutions, small insertions and deletions, and short tandem duplications. Classification under the ACMG/AMP 2015 framework depends heavily on the molecular consequence (null vs missense vs in-frame), so accurate variant typing is the first step in interpretation. See the ACMG Variant Classification leaf for the full criteria framework.

A single base substitution at one genomic position.

• Transitions vs transversions: transitions (purine to purine or pyrimidine to pyrimidine, A<>G and C<>T) outnumber transversions roughly 2:1 in the human genome despite there being twice as many possible transversions. CpG dinucleotides show especially high transition rates because methylated cytosine deaminates to thymine, making C>T (and G>A on the opposite strand) the most common substitution at CpG sites.
• Coding consequences:
  • Synonymous (silent): codon changes but amino acid is preserved. Usually benign, but can disrupt splicing or codon usage.
  • Missense: amino acid substitution. Conservative swaps preserve biochemical class (e.g., Leu>Ile); non-conservative swaps change charge or polarity (e.g., Arg>Cys) and are more likely damaging.
  • Nonsense: introduces a premature termination codon (PTC). PTCs more than ~50 bp upstream of the last exon-exon junction trigger nonsense-mediated decay (NMD) and produce a null allele (PVS1).
  • Start-loss: disrupts the initiator methionine; often null but depends on availability of downstream alternative start codons.
  • Stop-loss: read-through into the 3' UTR, producing C-terminally extended protein.
• Splice-site variants: canonical sites are the donor +1/+2 and acceptor -1/-2 dinucleotides; disruption almost always abolishes splicing and qualifies as PVS1 (null). Variants at extended splice sites (donor +3 to +6, acceptor -3 to -14, branch point) or deep intronic positions may create cryptic splice sites and require splicing predictors (SpliceAI, MaxEntScan) or RNA studies.

Small (1-50 bp) insertions or deletions, most commonly arising from replication slippage at repetitive tracts or microhomology-mediated end joining.

• In-frame indels (multiples of 3): preserve reading frame but alter protein by adding or removing residues. Coded as PM4 (protein length change in non-repeat region) under ACMG.
• Frameshift indels: shift the reading frame, producing a downstream PTC. Trigger NMD unless within the last exon or near the 3' end of the penultimate exon, in which case a truncated protein may escape NMD. Frameshifts in NMD-competent regions are PVS1.
• Microhomology (2-25 bp shared sequence flanking a deletion) is a hallmark of microhomology-mediated end joining and helps distinguish replication-error indels from larger structural rearrangements.

Short, in-frame duplications of one or more exons or codons that produce an elongated protein. Frequently disease-causing through gain-of-function or dominant-negative mechanisms.

• MECP2 duplication syndrome (Xq28): whole-gene duplication causes severe intellectual disability and recurrent infections in males.
• CDKL5 intragenic duplications: cause early-onset epileptic encephalopathy.
• Coded under PVS1 only when the duplication clearly disrupts the reading frame; otherwise PM4 or case-specific evidence applies.

Trinucleotide and other short tandem repeat expansions (CAG in HTT, CGG in FMR1, GAA in FXN, CTG in DMPK) cause distinct disease classes covered in the triplet expansion disorders content. Standard short-read NGS frequently misses or undercalls expansions because reads cannot span the full repeat tract. Detection requires:

• Repeat-primed PCR (RP-PCR), the historical gold standard for CGG and CAG.
• Southern blot for very large expansions (full mutation FMR1).
• NGS-based callers such as ExpansionHunter and STRetch on short-read data, or long-read sequencing (PacBio, ONT) for full-length resolution.

• PVS1: null variant in a gene where loss of function is a known mechanism (nonsense, frameshift, canonical +/-1,2 splice, multi-exon deletion, start-loss in some contexts).
• PM4: in-frame indel or stop-loss producing a protein length change.
• PP3 / BP4: computational evidence (REVEL, CADD, SpliceAI) for or against pathogenicity.
• PS1 / PM5: same amino acid change (PS1) or different change at the same residue (PM5) as a known pathogenic variant.