esl-mixdchlet(1)

1esl-mixdchlet(1)                 Easel Manual                 esl-mixdchlet(1)
2
3
4

NAME

6       esl-mixdchlet - fitting mixture Dirichlets to count data
7
8

SYNOPSIS

10       esl-mixdchlet fit [options] Q K in_countfile out_mixchlet
11         (train a new mixture Dirichlet)
12
13       esl-mixdchlet score [options] mixdchlet_file counts_file
14         (calculate log likelihood of count data, given mixture Dirichlet)
15
16       esl-mixdchlet gen [options] mixdchlet_file
17         (generate synthetic count data from mixture Dirichlet)
18
19       esl-mixdchlet sample [options]
20         (sample a random mixture Dirichlet for testing)
21
22
23

DESCRIPTION

25       The  esl-mixdchlet  miniapp  is  for training mixture Dirichlet priors,
26       such as the priors used in HMMER and Infernal. It has four subcommands:
27       fit,  score,  gen,  and  sample.  The most important subcommand is fit,
28       which is the subcommand for fitting a new mixture  Dirichlet  distribu‐
29       tion to a collection of count vectors (for example, emission or transi‐
30       tion count vectors from Pfam or Rfam training sets).
31
32
33       Specifically, esl-mixdchlet fit fits a new mixture Dirichlet  distribu‐
34       tion with Q mixture components to the count vectors (of alphabet size K
35       ) in input file in_countfile, and saves the mixture Dirichlet into out‐
36       put file out_mixdchlet.
37
38
39       The  input  count vector file in_countfile contains one count vector of
40       length K fields per line, for any number of  lines.   Blank  lines  and
41       lines  starting  in  #  (comments) are ignored.  Fields are nonnegative
42       real values; they do not have to  be  integers,  because  they  can  be
43       weighted counts.
44
45
46       The format of a mixture Dirichlet file out_mixdchlet is as follows. The
47       first line has two fields, K Q, where K is the alphabet size and  Q  is
48       the  number  of  mixture  components.   The next Q lines consist of K+1
49       fields. The first field is the mixture coefficient q_k, followed  by  K
50       fields with the Dirichlet alpha[k][a] parameters for this component.
51
52
53       The esl-mixdchlet score subcommand calculates the log likelihood of the
54       count vector data in counts_file, given the mixture Dirichlet in  mixd‐
55       chlet_file.
56
57
58       The  esl-mixdchlet gen subcommand generates synthetic count data, given
59       a mixture Dirichlet.
60
61
62       The esl-mixdchlet sample subcommand creates a random mixture  Dirichlet
63       distribution and outputs it to standard output.
64
65
66

OPTIONS FOR FIT SUBCOMMAND

68       -h     Print brief help specific to the fit subcommand.
69
70
71       -s <seed>
72              Set  random number generator seed to nonnegative integer <seed>.
73              Default is 0, which means to use a quasirandom  arbitrary  seed.
74              Values >0 give reproducible results.
75
76
77
78
79

OPTIONS FOR SCORE SUBCOMMAND

81       -h     Print brief help specific to the score subcommand.
82
83
84
85

OPTIONS FOR GEN SUBCOMMAND

87       -h     Print brief help specific to the gen subcommand.
88
89
90       -s <seed>
91              Set  random number generator seed to nonnegative integer <seed>.
92              Default is 0, which means to use a quasirandom  arbitrary  seed.
93              Values >0 give reproducible results.
94
95
96
97       -M <M> Generate <M> counts per sampled vector. (Default 100.)
98
99
100       -N <N> Generate <N> count vectors. (Default 1000.)
101
102
103

OPTIONS FOR SAMPLE SUBCOMMAND

105       -h     Print brief help specific to the sample subcommand.
106
107
108       -s <seed>
109              Set  random number generator seed to nonnegative integer <seed>.
110              Default is 0, which means to use a quasirandom  arbitrary  seed.
111              Values >0 give reproducible results.
112
113
114
115       -K <K> Set the alphabet size to <K>.  (Default is 20, for amino acids.)
116
117
118       -Q <Q> Set the number of mixture components to <Q>.  (Default is 9.)
119
120
121
122
123

COPYRIGHT

129       Copyright (C) 2020 Howard Hughes Medical Institute.
130       Freely distributed under the BSD open source license.
131
132

AUTHOR

134       http://eddylab.org
135
136
137
138
139Easel 0.48                         Nov 2020                   esl-mixdchlet(1)