discovering new subsets of stereotyped antigen receptor sequences
lab home: http://bat.infspire.org
citation: http://www.ncbi.nlm.nih.gov/pubmed/19759557
code: we provide Perl scripts,
IBM's TEIRESIAS (old version) and NR-grep by G. Navarro for Ubuntu 64bit,
and demo data
usage:
teiresias.pl--in -> table: tab-delimited, NO header line, Unix ASCII - see demo.table 1st column: required! label, make sure they're unique 2nd column: required! amino acid junction sequence, with anchors 3rd column: optional IMGT-formatted V gene (and allele) 4th column: optional annotation
other (optional) arguments - defaults are current standards for CLL (PMID:22415752) --ide -> minimum identity [0-1], default: 0.5 --sim -> minimum similarity [0-1], default: 0.7 --lendiff -> max length difference [0,1,2,3], default: 0 --offdiff -> max offset difference [0,1,2,3], default: 0 --nophylo -> ignore phylogenetic relationships as defined in 'phyloequiv' --demo -> run demo (~30sec in a multicore system) --debug -> do not erase intermediate files
notes / hints / tips TEIRESIAS and NR-grep are OK on Ubuntu 64bit, otherwise download/compile from original sources (be aware though that the latest TEIRESIAS version (0.9.1, 3/13/2014) gives unexpected results). When possible, sequences in results table are uppercased with the pattern that clustered them. You can find the same sequence in different clusters (i.e. ~fuzzy clustering); these are used as linkers for higher-level clustering (L1, L2, etc), where linkage can actually be very loose so be careful with interpretation.