Skip to content

HGVS Nomenclature Grammar#

The HGVS Nomenclature is fundamentally a computational language, but it has historically been written using human-interpreted descriptions rather than using a precise formal grammar. While migrating from human descriptions to a formal grammar would increase formal precision, it would be disruptive to the community. Instead, the HVNC has chosen to selectively adopt principles and presentation of computational grammars in order to improve the precision of definitions while preserving human readability. We have chosen a common variation of the Extended Backus-Naur Form (EBNF) to define variant syntax.

We recognize that most users of HGVS will not want to concern themselves with the pedagogy of formal grammars and that most discussions of the topic are dense. This primer describes the most essential concepts for EBNF as used by the HGVS Nomenclature, presented slightly informally to make them accessible to a broad audience.


  • A grammar specifies the valid syntax for a string of characters. A symbol is a grammatical classification that aids parsing.
  • There are two kinds of symbols: a literal symbol represents verbatim text, and a non-terminal symbol represents a named concept like "an integer sequence position". more
  • Literal symbols are specified in double quotes. For example, `"A"`` means that an "A" character is accepted as input (and nothing else).
  • A "pipe" (|) separates alternatives. For example "A" | "B" accepts a single "A" or "B" character (and nothing else).
  • Symbols (of either type) may be concatenated. For example, digit digit matches a two-digit number. (Note: HGVS Nomenclature uses a common variant of EBNF in which concatenation is implied by adjacent symbols. Other EBNF variants require a comma between concatenated symbols.)
  • A production rule defines a non-terminal symbol as a pattern of other symbols. For example, digit = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9" says that digit is made of "0" or "1" or "2", etc. Similarly, two_digit_dumber = digit digit defines a two_digit_number non-terminal symbol as being composed of two adjacent digits.
  • Modfiers are used to indicate cardinality: * indicates zero or more of the preceeding rule or symbol; + indicates one or more; and ? denotes zero or one (i.e., optional). For example, number = digit+ recognizes any sequence of digits without sign or decimal point.
  • Parenthesis may be used to group a subexpression. For example, "A" ( "B" | "C" ) accepts "AB" or "AC", and nothing else.
  • Square brackets indicate optional syntax. For example, sign = "+" | "-" and number = [sign] digit+ allow number to defined as any sequence of digets, optionally with a single prefix sign.

Putting this all together, a rule to match NCBI RefSeq accessions might be written as refseq_accession = letter letter "_" digit+ "." digit+, which would accept accessions like "NM_000551.3" and "NC_000012.12". Note that, as defined here, refseq_accession would also accept invalid or unlikely accessions such as "QQ_012.3333333333333333"; this example highlights the distinction between syntactic and semantic validity.