|
kkonganti@11
|
1 // Help text for mash sketch within CPIPES.
|
|
kkonganti@11
|
2
|
|
kkonganti@11
|
3 def mashsketchHelp(params) {
|
|
kkonganti@11
|
4
|
|
kkonganti@11
|
5 Map tool = [:]
|
|
kkonganti@11
|
6 Map toolspecs = [:]
|
|
kkonganti@11
|
7 tool.text = [:]
|
|
kkonganti@11
|
8 tool.helpparams = [:]
|
|
kkonganti@11
|
9
|
|
kkonganti@11
|
10 toolspecs = [
|
|
kkonganti@11
|
11 'mashsketch_run': [
|
|
kkonganti@11
|
12 clihelp: 'Run `mash screen` tool. Default: ' +
|
|
kkonganti@11
|
13 (params.mashsketch_run ?: false),
|
|
kkonganti@11
|
14 cliflag: null,
|
|
kkonganti@11
|
15 clivalue: null
|
|
kkonganti@11
|
16 ],
|
|
kkonganti@11
|
17 'mashsketch_l': [
|
|
kkonganti@11
|
18 clihelp: 'List input. Lines in each <input> specify paths to sequence files, ' +
|
|
kkonganti@11
|
19 'one per line. ' +
|
|
kkonganti@11
|
20 "Default: ${params.mashsketch_l}",
|
|
kkonganti@11
|
21 cliflag: '-l',
|
|
kkonganti@11
|
22 clivalue: (params.mashsketch_l ? ' ' : '')
|
|
kkonganti@11
|
23 ],
|
|
kkonganti@11
|
24 'mashsketch_I': [
|
|
kkonganti@11
|
25 clihelp: '<path> ID field for sketch of reads (instead of first sequence ID). ' +
|
|
kkonganti@11
|
26 "Default: ${params.mashsketch_I}",
|
|
kkonganti@11
|
27 cliflag: '-I',
|
|
kkonganti@11
|
28 clivalue: (params.mashsketch_I ?: '')
|
|
kkonganti@11
|
29 ],
|
|
kkonganti@11
|
30 'mashsketch_C': [
|
|
kkonganti@11
|
31 clihelp: '<path> Comment for a sketch of reads (instead of first sequence comment). ' +
|
|
kkonganti@11
|
32 "Default: ${params.mashsketch_C}",
|
|
kkonganti@11
|
33 cliflag: '-C',
|
|
kkonganti@11
|
34 clivalue: (params.mashsketch_C ?: '')
|
|
kkonganti@11
|
35 ],
|
|
kkonganti@11
|
36 'mashsketch_k': [
|
|
kkonganti@11
|
37 clihelp: '<int> K-mer size. Hashes will be based on strings of this many ' +
|
|
kkonganti@11
|
38 'nucleotides. Canonical nucleotides are used by default (see ' +
|
|
kkonganti@11
|
39 'Alphabet options below). (1-32) ' +
|
|
kkonganti@11
|
40 "Default: ${params.mashsketch_k}",
|
|
kkonganti@11
|
41 cliflag: '-k',
|
|
kkonganti@11
|
42 clivalue: (params.mashsketch_k ?: '')
|
|
kkonganti@11
|
43 ],
|
|
kkonganti@11
|
44 'mashsketch_s': [
|
|
kkonganti@11
|
45 clihelp: '<int> Sketch size. Each sketch will have at most this many non-redundant ' +
|
|
kkonganti@11
|
46 'min-hashes. ' +
|
|
kkonganti@11
|
47 "Default: ${params.mashsketch_s}",
|
|
kkonganti@11
|
48 cliflag: '-s',
|
|
kkonganti@11
|
49 clivalue: (params.mashsketch_s ?: '')
|
|
kkonganti@11
|
50 ],
|
|
kkonganti@11
|
51 'mashsketch_i': [
|
|
kkonganti@11
|
52 clihelp: 'Sketch individual sequences, rather than whole files, e.g. for ' +
|
|
kkonganti@11
|
53 'multi-fastas of single-chromosome genomes or pair-wise gene ' +
|
|
kkonganti@11
|
54 'comparisons. ' +
|
|
kkonganti@11
|
55 "Default: ${params.mashsketch_i}",
|
|
kkonganti@11
|
56 cliflag: '-i',
|
|
kkonganti@11
|
57 clivalue: (params.mashsketch_i ? ' ' : '')
|
|
kkonganti@11
|
58 ],
|
|
kkonganti@11
|
59 'mashsketch_S': [
|
|
kkonganti@11
|
60 clihelp: '<int> Seed to provide to the hash function. (0-4294967296) [42] ' +
|
|
kkonganti@11
|
61 "Default: ${params.mashsketch_S}",
|
|
kkonganti@11
|
62 cliflag: '-S',
|
|
kkonganti@11
|
63 clivalue: (params.mashsketch_S ?: '')
|
|
kkonganti@11
|
64 ],
|
|
kkonganti@11
|
65
|
|
kkonganti@11
|
66 'mashsketch_w': [
|
|
kkonganti@11
|
67 clihelp: '<num> Probability threshold for warning about low k-mer size. (0-1) ' +
|
|
kkonganti@11
|
68 "Default: ${params.mashsketch_w}",
|
|
kkonganti@11
|
69 cliflag: '-w',
|
|
kkonganti@11
|
70 clivalue: (params.mashsketch_w ?: '')
|
|
kkonganti@11
|
71 ],
|
|
kkonganti@11
|
72 'mashsketch_r': [
|
|
kkonganti@11
|
73 clihelp: 'Input is a read set. See Reads options below. Incompatible with ' +
|
|
kkonganti@11
|
74 '--mashsketch_i. ' +
|
|
kkonganti@11
|
75 "Default: ${params.mashsketch_r}",
|
|
kkonganti@11
|
76 cliflag: '-r',
|
|
kkonganti@11
|
77 clivalue: (params.mashsketch_r ? ' ' : '')
|
|
kkonganti@11
|
78 ],
|
|
kkonganti@11
|
79 'mashsketch_b': [
|
|
kkonganti@11
|
80 clihelp: '<size> Use a Bloom filter of this size (raw bytes or with K/M/G/T) to ' +
|
|
kkonganti@11
|
81 'filter out unique k-mers. This is useful if exact filtering with ' +
|
|
kkonganti@11
|
82 '--mashsketch_m uses too much memory. However, some unique k-mers may pass ' +
|
|
kkonganti@11
|
83 'erroneously, and copies cannot be counted beyond 2. Implies --mashsketch_r. ' +
|
|
kkonganti@11
|
84 "Default: ${params.mashsketch_b}",
|
|
kkonganti@11
|
85 cliflag: '-b',
|
|
kkonganti@11
|
86 clivalue: (params.mashsketch_b ?: '')
|
|
kkonganti@11
|
87 ],
|
|
kkonganti@11
|
88 'mashsketch_m': [
|
|
kkonganti@11
|
89 clihelp: '<int> Minimum copies of each k-mer required to pass noise filter for ' +
|
|
kkonganti@11
|
90 'reads. Implies --mashsketch_r. ' +
|
|
kkonganti@11
|
91 "Default: ${params.mashsketch_r}",
|
|
kkonganti@11
|
92 cliflag: '-m',
|
|
kkonganti@11
|
93 clivalue: (params.mashsketch_m ?: '')
|
|
kkonganti@11
|
94 ],
|
|
kkonganti@11
|
95 'mashsketch_c': [
|
|
kkonganti@11
|
96 clihelp: '<num> Target coverage. Sketching will conclude if this coverage is ' +
|
|
kkonganti@11
|
97 'reached before the end of the input file (estimated by average ' +
|
|
kkonganti@11
|
98 'k-mer multiplicity). Implies --mashsketch_r. ' +
|
|
kkonganti@11
|
99 "Default: ${params.mashsketch_c}",
|
|
kkonganti@11
|
100 cliflag: '-c',
|
|
kkonganti@11
|
101 clivalue: (params.mashsketch_c ?: '')
|
|
kkonganti@11
|
102 ],
|
|
kkonganti@11
|
103 'mashsketch_g': [
|
|
kkonganti@11
|
104 clihelp: '<size> Genome size (raw bases or with K/M/G/T). If specified, will be used ' +
|
|
kkonganti@11
|
105 'for p-value calculation instead of an estimated size from k-mer ' +
|
|
kkonganti@11
|
106 'content. Implies --mashsketch_r. ' +
|
|
kkonganti@11
|
107 "Default: ${params.mashsketch_g}",
|
|
kkonganti@11
|
108 cliflag: '-g',
|
|
kkonganti@11
|
109 clivalue: (params.mashsketch_g ?: '')
|
|
kkonganti@11
|
110 ],
|
|
kkonganti@11
|
111 'mashsketch_n': [
|
|
kkonganti@11
|
112 clihelp: 'Preserve strand (by default, strand is ignored by using canonical ' +
|
|
kkonganti@11
|
113 'DNA k-mers, which are alphabetical minima of forward-reverse ' +
|
|
kkonganti@11
|
114 'pairs). Implied if an alphabet is specified with --mashsketch_a ' +
|
|
kkonganti@11
|
115 'or --mashsketch_z. ' +
|
|
kkonganti@11
|
116 "Default: ${params.mashsketch_n}",
|
|
kkonganti@11
|
117 cliflag: '-n',
|
|
kkonganti@11
|
118 clivalue: (params.mashsketch_n ? ' ' : '')
|
|
kkonganti@11
|
119 ],
|
|
kkonganti@11
|
120 'mashsketch_a': [
|
|
kkonganti@11
|
121 clihelp: 'Use amino acid alphabet (A-Z, except BJOUXZ). Implies ' +
|
|
kkonganti@11
|
122 '--mashsketch_n --mashsketch_k 9. ' +
|
|
kkonganti@11
|
123 "Default: ${params.mashsketch_a}",
|
|
kkonganti@11
|
124 cliflag: '-a',
|
|
kkonganti@11
|
125 clivalue: (params.mashsketch_a ? ' ' : '')
|
|
kkonganti@11
|
126 ],
|
|
kkonganti@11
|
127 'mashsketch_z': [
|
|
kkonganti@11
|
128 clihelp: '<text> Alphabet to base hashes on (case ignored by default; ' +
|
|
kkonganti@11
|
129 'see --mashsketch_Z). K-mers with other characters will be ' +
|
|
kkonganti@11
|
130 'ignored. Implies --mashsketch_n. ' +
|
|
kkonganti@11
|
131 "Default: ${params.mashsketch_z}",
|
|
kkonganti@11
|
132 cliflag: '-z',
|
|
kkonganti@11
|
133 clivalue: (params.mashsketch_z ?: '')
|
|
kkonganti@11
|
134 ],
|
|
kkonganti@11
|
135 'mashsketch_Z': [
|
|
kkonganti@11
|
136 clihelp: 'Preserve case in k-mers and alphabet (case is ignored by default). ' +
|
|
kkonganti@11
|
137 'Sequence letters whose case is not in the current alphabet will be ' +
|
|
kkonganti@11
|
138 'skipped when sketching. ' +
|
|
kkonganti@11
|
139 "Default: ${params.mashsketch_Z}",
|
|
kkonganti@11
|
140 cliflag: '-Z',
|
|
kkonganti@11
|
141 clivalue: (params.mashsketch_Z ?: '')
|
|
kkonganti@11
|
142 ]
|
|
kkonganti@11
|
143 ]
|
|
kkonganti@11
|
144
|
|
kkonganti@11
|
145 toolspecs.each {
|
|
kkonganti@11
|
146 k, v -> tool.text['--' + k] = "${v.clihelp}"
|
|
kkonganti@11
|
147 tool.helpparams[k] = [ cliflag: "${v.cliflag}", clivalue: v.clivalue ]
|
|
kkonganti@11
|
148 }
|
|
kkonganti@11
|
149
|
|
kkonganti@11
|
150 return tool
|
|
kkonganti@11
|
151 } |
