Skip to content

Latest commit

 

History

History
59 lines (47 loc) · 3.71 KB

README.md

File metadata and controls

59 lines (47 loc) · 3.71 KB

The Great Cattle Antibody Repertoire

Longitudinal study of vaccination of 204 calves against the bovine respiratory disease

Repository files

  • aux: antibody titers, calf ages, V gene usages, ultralong CDRH3 fractions;
  • gsv_finding: intermediate results of IGHV gene genotyping, detected germline and somatic variations (GSVs) and their R-ratios;
  • igqtl_results: clusters of calves computed using R-ratios of the detected GSVs.

AIRR metadata

Metadata for 204 processed antibody repertoires are available here.

Each zipped CSV annotation file describes four Rep-Seq libraries corresponding to a single animal and includes all distinct VDJ nucleotide sequences. Each line represents a distinct nucleotide sequence and contains with the following information:

  • Nucleotide and amino acid sequences,
  • Read count,
  • Time point (–3, 0, 3, or 6),
  • IGHV and IGHJ gene hits,
  • Nucleotide and amino acid sequences of CDR1, CDR2, CDR3,
  • Whether or not CDR3 is ultralong (i.e., the length of CDR3 exceeds 150 nt),
  • Clone ID. Clones are specific to time point, i.e., they are computed for sequences from a given time point. A clone is defined as a group of sequences sharing V hit and J hit and having identical CDR3s.
  • Alignment of the variable gene segment against the closest germline IGHV gene.

Germline and somatic variation data

  • vgene_genotyping: the directory contains subdirectories corresponding to 204 cattle subjects. Each subdirectory consists of TXT files with nucleotide counts for all positions of all V genes. Each file corresponds to a single V gene and contains a table showing the numbers of clones (in this case, clones = sequences sharing V and J genes and having identical CDR3s) with A, C, G, and T nucleotides at positions 0,...,L-1, where L is the length of the V gene. A fragment of the 14007_mismatches/IGHV1-7_mismatches.txt file is shown below. E.g., 4, 7, 10010, and 10 clones have A, C, G, and T at position 38 (0-based) of IGHV1-7, respectively.
Position NumAs NumCs NumGs NumTs
...
38 4 7 10010 10
39 1 9978 2 50
40 0 10018 6 7
41 24 3664 6332 11
42 5 24 7 9995
...
  • gsv_r_ratios.txt: TXT files containing R-ratios for 52 GSVs detected by IgQTL tool using the nucleotide counts from vgene_genotyping.

Scripts

  • igqtl.py: takes vgene_genotyping directory as an input, computes germline and somatic variations, and reports their R-ratios for all subjects:
python igqtl.py vgene_genotyping_dir output_dir

Output directory contains R-ratios of GSVs per V gene and all GSV combined together. Each identified GSV is written as a column and have following format: VGENE:POSITION_N1N2, where POSITION is 0-based, N1 and N2 are the most abundant and the second most abundant nucleotides at this position, respectively.

  • compute_pca_clusters.py: takes the file with R-ratios of GSVs and compute clusters of subjects using PCA:
python compute_pca_clusters.py gsv_r_ratios.txt output_dir

Python dependencies:

  • scipy, matplotlib, numpy (usually a part of the standard python installation with conda)
  • kneed (conda install -c conda-forge kneed)
  • pandas (conda install -c anaconda pandas)
  • seaborn (conda install -c anaconda seaborn)
  • sklearn (conda install -c anaconda scikit-learn)

Reference:

Safonova Y, Shin SB, Kramer L, Reecy J, Watson CT, Smith TPL, Pevzner PA, Variations in antibody repertoires correlate with vaccine responses. Accepted to Genome Research, 2021 [preprint].