Skip to content

Symbol tables

Every byte of a sequence line belongs to exactly one class, which depends on the alphabet. The classes are the values of class in run.validation.characters of the results. Letters are case-insensitive: lower-case letters are folded to upper case first (and counted).

ClassEffect at a position
residuepart of the k-mer
gap, missing, ambiguity, stop, frameshiftthe k-mer is invalid: that sequence does not count towards the support
whitespaceignored (the sequence continues)
unknownstrict mode: the run stops with record, line and column; permissive and report modes: the k-mer is invalid and the character is counted and reported
ClassCharactersSource
residueA C D E F G H I K L M N P Q R S T V W Ythe 20 standard amino acids
residueU (selenocysteine), O (pyrrolysine)genetically encoded amino acids (Böck et al. 1991; Srinivasan et al. 2002)
ambiguityX (any), B (D or N), Z (E or Q), J (I or L)IUPAC-IUB one-letter codes
gap- . _alignment gaps (Eddy 2011)
missing~ ?missing data (Eddy 2011); ? is “no consensus” in LANL alignments (Los Alamos HIV Sequence Database 2009)
stop* $ ^* is the usual stop symbol of translations; $ and ^ are stops in LANL alignments (Los Alamos HIV Sequence Database 2009; Los Alamos HIV Sequence Database 2024)
frameshift#LANL alignments (Los Alamos HIV Sequence Database 2009; Los Alamos HIV Sequence Database 2024)
whitespacespace, tab (\t), carriage return (\r), vertical tab, form feedline formatting; Easel’s FASTA reader ignores the first three in sequence lines (Eddy 2011)
unknownevery other byte: digits, other punctuation, the letters not listed above, non-ASCII bytes
ClassCharactersSource
residueA C G T
residueU, counted as TT and U are one symbol; the number folded is reported
ambiguityR Y K M S W B D H V NIUPAC-IUB nucleotide codes (Cornish-Bowden 1985)
ambiguityX, I (inosine)“any” and inosine, which pairs with several bases (Martin et al. 1985; Ohtsuka et al. 1985)
gap, missing, stop, frameshift, whitespaceas for proteins
unknownevery other byte

As DNA, with U as the residue and T counted as U.

  • The line feed ends a line; a carriage return directly before it is part of the line ending.
  • A NUL byte (0x00) anywhere stops the run in every mode: the file is binary or corrupt.
  • Header lines (starting with >) are not sequence data; see header metadata.

Folding case, treating T and U as one symbol, excluding stops and frameshifts, and stopping at unknown characters by default are deliberate; the reasons, and the effect on the paper’s HIV-1 pol protein dataset, are in Reading alignments. DiMA v5.0.9 used different rules; see Differences from DiMA v5.0.9.