Appendix

TCR/BCR refenrece sequences library

Default list and sequences of V, D, J and C genes used by MiXCR are taken from GenBank. Accession numbers of records used for each locus are listed in the following table:

     

Homo sapiens

 TRA/TRD NG_001332.2
TRB NG_001333.2
TRG NG_001336.2
IGH NG_001019.5
IGK NG_000834.1
IGL NG_000002.1
     

Mus musculus

 TRA/TRD NG_007044.1
TRB NG_006980.1
TRG NG_007033.1
IGH NG_005838.1
IGK NG_005612.1
IGL NG_004051.1

Alignment and mutations encoding

MiXCR outputs alignments in exportClones and exportAlignments as a list of 7 fields separated by | symbol as follows:

targetFrom | targetTo | targetLength | queryFrom | queryTo | mutations | alignmentScore

where

  • targetFrom - position of first aligned nucleotide in target sequence (sequence of gene feature from reference V, D, J or C gene used in alignment; e.g. VRegion in TRBV12-2); this boundary is inclusive
  • targetTo - next position after last aligned nucleotide in target sequence; this boundary is exclusive
  • targetLength - length of target sequence (e.g. length of VRegion in TRBV12-2)
  • queryFrom - position of first aligned nucleotide in query sequence (sequence of sequencing read or clonal sequence); this boundary is inclusive
  • queryTo - next position after last aligned nucleotide in query sequence; this boundary is exclusive
  • mutations - list of mutations from target sequence to query sequence (see below)
  • alignmentScore - score of alignment

all positions are zero-based (i.e. first nucleotide has index 0)

Mutations are encoded as a list of single-nucleotide edits (similar to what is used in definition of Levenshtein distance, i.e. insertions, deletions or substitutions); if one apply these mutations to aligned subsequence of target sequence, one will obtain aligned subsequence of query sequence.

Each single mutation (single-nucleotide edit) is encoded in the following way (without any spaces; some fields may absent in some cases, see description):

type [fromNucleotide] position [toNucleotide]

  • type of mutation (one letter):
  • S for substitution
  • D for deletion
  • I for insertion
  • fromNucleotide is a nucleotide in target sequence affected by mutation (applicable only for substitutions and deletions; absent for insertions)
  • position is a zero-based absolute position in target sequence affected by mutation; for insertions denotes position in target sequence right after inserted nucleotide
  • toNucleotide nucleotide after mutation (applicable only for substitutions and insertions; absent for deletions)

Note, that for deletions and substitutions

targetSequence[position] == fromNucleotide

i.e. target sequence always have fromNucleotide at position position; for insertions fromNucleotide field is absent

Here are several examples of single mutations:

  • SA4T - substitution of A at position 4 to T
  • DC12 - deletion of C at position 12
  • I15G - insertion of G before position 15

Consider the following BLAST-like alignments encoded in MiXCR notation:

  • Alignment without mutation

    target = TTGTGCTGACAGATACCCC
query  = CGAGTGCTGACAGATACCGTCGATGCT

BLAST like alignment:
2 GTGCTGACAGATACC 16
  |||||||||||||||
3 GTGCTGACAGATACC 17

MiXCR alignment:
0|15|17|3|18||75.0

subsequence from target (from nucleotide 0 to nucleotide 15) was found to be identical to susequence from query (from nucleotide 3 to nucleotide 18).

  • Alignment with mutation

    target = TTGTGCTGACAGATACCCC
query  = CGAGTGCTATAGACTACCGTCGATGCT

BLAST like alignment:
2 GTGCTGACAGA-TACC 16
  ||||| | ||| ||||
3 GTGCT-ATAGACTACC 17

MiXCR alignment:
0|15|17|3|18|DG7SC9TI13C|41.0

so, to obtain subseqeunce from query sequence from 3 to 18 we need to apply the following mutations to subsequence of target sequence from 2 to 16: - deletion of G at position 7 - substitution of C at position 9 to T - insertion of C before at position 13