Skip to content

README.md

Latest commit

 

History

History
54 lines (45 loc) · 3.88 KB

README.md

File metadata and controls

54 lines (45 loc) · 3.88 KB
Raw

Updating a germline set (general recipe)

Use merge-germline-set.py to merge a freshly-downloaded set into an existing one. It's always a union (we never drop a gene we already had). The two modes differ in which name wins when both sets have the same sequence under different names:

  • augmentation: the new genes are from a different source (e.g. engelbrecht/rodriguez on top of imgt), so we keep our established (old) names.
  • replacement: the new set is a newer version of the same source (e.g. a fresh ogrdb pull), so it's authoritative -- its (new) names win, and any old genes missing from it get a warning (maybe real upstream removals, maybe a botched download).

augmentation (new genes from a different source), new set already in partis layout:

./data/germlines/merge-germline-set.py --species human --new-dir --expected augmentation --write

replacement (newer version of the same source):

./data/germlines/merge-germline-set.py --species macaque --new-dir data/germlines/ogrdb-download/macaque --expected replacement --write

--new-dir must be in partis layout (<dir>/<locus>/<locus>{v,d,j}.fasta [+ extras.csv]). For an ogrdb download that is a single combined fasta, first split it per region/locus (see the ogrdb download + split notes below). Codon positions for new genes are inferred by aligning to the old set, so an extras.csv in the new set is optional; add --sanitize-names only if the new set's fasta ids aren't already valid gene names. Default is a dry run; add --write to write the merged set. It prints the added and old-only gene names plus the conflicts (redirect stdout to a file for a record -- see the per-species changelog/ dirs). Review the output and every warning before committing.

human: download from https://ogrdb.airr-community.org/germline_sets/Human

  • use the "AIRR-seq reference set" (not the "source set" -- this has duplicates, among other things)
  • make a dir, separate v/d/j files
  • run (separately for each locus): ./bin/partis cache-parameters --locus igh --infname test/mishmash.fa --n-max-queries 1 --sanitize-input-germlines --initial-germline-dir
    • input file doesn't matter/isn't used
    • add something like: glutils.write_glfo('ogrdb-ref/ogrdb-sanitized', glfo, debug=True) to bin/partis where args.sanitize_input_germlines gets used
  • also added genes from genomic sequencing in engelbrecht/rodriguez papers, downloaded from vdjbase

macaque: from ogrdb https://ogrdb.airr-community.org/germline_sets/Macaca%20mulatta

mouse/ is merged from imgt, ogrdb c57bl, and ogrdb balbc

c57bl/ and balbc/ are just from ogrdb

to update mouse germlines, download from: https://ogrdb.airr-community.org/germline_sets/Mouse

  1. cp unaligned fastas to appropriate dir in ogrdb-downloads/
  2. split apart igh files, cp igk/igl files from germlines/mouse/
  3. merge in with merge-germline-set.py (see "Updating a germline set" above), once per mouse type
  4. after checking that everything went well, replace old mouse/ dirs with new merged dirs

leader and constant genes from imgt, with extra c genes from the flairr-seq authors

  • Any allele with a name of FL or *N is novel and not found in the IMGT database.
  • Any allele with the name -FL or *_ext is an extension of a documented IMGT allele. Some alleles with the first syntax have a letter signifying it is an extension of an IMGT allele but contains unique sequence from other extensions of that allele.
  • i.e. there's tons of duplicate stuff in here and i'm not sorting through it, whatever