File: subsample.txt

package info (click to toggle)
bali-phy 4.0-1
  • links: PTS, VCS
  • area: main
  • in suites: trixie
  • size: 15,392 kB
  • sloc: cpp: 120,442; xml: 13,966; haskell: 9,975; python: 2,936; yacc: 1,328; perl: 1,169; lex: 912; sh: 343; makefile: 26
file content (27 lines) | stat: -rw-r--r-- 878 bytes parent folder | download | duplicates (7)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
# The `--subsample` command:

-x          <number>                        Factor by which to subsample.
--subsample <number>

This option should usually not be used.  BAli-Phy performs more
work per iteration than some other software, so each sample
contains new information.  Subsampling is also mathematically
never necessary for correctness, and it can decrease accuracy
by throwing away information.  However, subsampling is sometimes
necessary to save disk space or processing time for large log files.

While numeric samples (`C1.log`) and tree sample (`C1.trees`) are
written every <number> iterations, alignments (`C1.P1.fastas`)
are written every <number>*10 iterations to save disk space.

# Examples:

   # Keep every 3rd sample in C1.log
   bali-phy dna.fasta -x3

   # Keep every sample in C1.log (the default)
   bali-phy dna.fasta -x1

# See also:

   iterations