Scroll to navigation



centrifuge-build - building index on all complete bacterial and viral genomes


centrifuge-build [options]* --conversion-table <table file> --taxonomy-tree <taxonomy tree file> <reference_in> <cf_index_base>


Centrifuge version 1.0.3-beta by Daehwan Kim (,

comma-separated list of files with ref sequences
write cf data to files with this dir/basename


reference sequences given on cmd line (as <reference_in>)
force generated index to be 'large', even if ref has fewer than 4 billion nucleotides
disable automatic -p/--bmax/--dcv memory-fitting
max bucket sz for blockwise suffix-array builder
max bucket sz as divisor of ref len (default: 4)
diff-cover period for blockwise (default: 1024)
disable diff-cover (algorithm becomes quadratic)
don't build .3/.4.bt2 (packed reference) portion
just build .3/.4.bt2 (packed reference) portion
SA is sampled every 2^offRate BWT chars (default: 5)
# of chars consumed in initial lookup (default: 10)
a table that converts any id to a taxonomy id
<file name> taxonomy tree
<file name> names corresponding to taxonomic IDs
<file name> table of contig (or genome) sizes
seed for random number generator
verbose output (for debugging)
number of alignment threads to launch (1)
k size for counting the number of distinct k-mer
print detailed description of tool and its options
print this usage message
print version information and quit


This manpage was written by Andreas Tille for the Debian distribution and can be used for any other usage of the program.

October 2021 centrifuge-build 1.0.3