Mercurial > repos > iuc > fasta_stats

comparison fasta-stats.xml @ 4:60718157dd5f draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/fasta_stats/ commit 50f5cce5a8c11001e2c59600a2b99a4243b6d06f"
author iuc
date Thu, 18 Nov 2021 20:56:27 +0000
parents 5b072a9eaa9d
children
comparison
equal deleted inserted replaced
3:5b072a9eaa9d 4:60718157dd5f
1 <tool id="fasta-stats" name="Fasta Statistics" version="1.0.3"> 1 <tool id="fasta-stats" name="Fasta Statistics" version="2.0" profile="20.05">
2 <description>Display summary statistics for a fasta file.</description> 2 <description>display summary statistics for a FASTA file</description>
3 <requirements> 3 <requirements>
4 <requirement type="package" version="5.26">perl</requirement> 4 <requirement type="package" version="1.21.4">numpy</requirement>
5 <requirement type="package" version="1.79">biopython</requirement>
5 </requirements> 6 </requirements>
6 <command detect_errors="exit_code"><![CDATA[ 7 <command detect_errors="exit_code"><![CDATA[
7 perl '${__tool_directory__}/fasta-stats.pl' 8 python '${__tool_directory__}/fasta-stats.py'
8 '$dataset' 9 --fasta '$fasta'
9 #if $genome_size: 10 --stats_output '$stats_output'
10 $genome_size 11 #if $gaps_option
12 --gaps_output '$gaps_output'
11 #end if 13 #end if
12 > '$stats' 14 #if $genome_size
15 --genome_size $genome_size
16 #end if
13 ]]> 17 ]]>
14 </command> 18 </command>
15 <inputs> 19 <inputs>
16 <param name="dataset" type="data" format="fasta" label="fasta or multifasta file" help="fasta dataset to get statistics for."/> 20 <param argument="--fasta" type="data" format="fasta" label="FASTA or Multi-FASTA file" help="FASTA dataset to get statistics."/>
17 <param name="genome_size" type="float" optional="True" label="Genome size estimate (optional)" help="Estimate of the genome size in bases. If specified, NG50 and LG50 will be calculated."/> 21 <param argument="--genome_size" type="integer" min="0" optional="true" label="Estimated genome size" help="This parameter is optional. If provided, it will be used for calculating the NG50 statistic." />
22 <param argument="--gaps_option" type="boolean" truevalue="true" falsevalue="false" label="Generate gap stats"/>
18 </inputs> 23 </inputs>
19 <outputs> 24 <outputs>
20 <data name="stats" format="tabular" label="${tool.name} on ${on_string}: Fasta summary stats"/> 25 <data name="stats_output" format="tabular" label="${tool.name} on ${on_string}: summary stats"/>
26 <data name="gaps_output" format="bed" label="${tool.name} on ${on_string}: Gap stats">
27 <filter>gaps_option</filter>
28 </data>
21 </outputs> 29 </outputs>
22 <tests> 30 <tests>
23 <test> 31 <test expect_num_outputs="1">
24 <param name="dataset" value="test.fasta"/> 32 <param name="fasta" value="test.fasta" ftype="fasta"/>
25 <output name="stats" file="test_out.txt"/> 33 <output name="stats_output" file="test_01.tab" ftype="tabular"/>
26 </test> 34 </test>
27 <test> 35 <!--Test gap options and NG50-->
28 <param name="dataset" value="ng50_input.fasta"/> 36 <test expect_num_outputs="2">
37 <param name="fasta" value="ng50_input.fasta" ftype="fasta"/>
38 <param name="gaps_option" value="true"/>
29 <param name="genome_size" value="4000"/> 39 <param name="genome_size" value="4000"/>
30 <output name="stats" file="ng50_out.txt"/> 40 <output name="stats_output" file="test_02.tab" ftype="tabular"/>
41 <output name="gaps_output" file="test_02.bed" ftype="bed"/>
42 </test>
43 <!--Compare outputs with QUAST-->
44 <test expect_num_outputs="1">
45 <param name="fasta" value="test_long_sequence.fasta" ftype="fasta"/>
46 <output name="stats_output" ftype="tabular">
47 <assert_contents>
48 <has_text text="8353"/>
49 <has_text text="303889"/>
50 <has_text text="22107"/>
51 </assert_contents>
52 </output>
31 </test> 53 </test>
32 </tests> 54 </tests>
33 <help> 55 <help><![CDATA[
34 **Fasta Stats** 56
35 Displays the summary statistics for a fasta file. 57 .. class:: infomark
58
59 **Purpose**
60
61 Displays the summary statistics for a FASTA file.
36 62
37 ------ 63 ------
38 64
39 Outputs in tabular form: 65 .. class:: infomark
40 Lengths: n50, min, max, median and average
41 66
42 Number of base pairs: A, C, G, T, N, Total and Total_not_N 67 **Outputs**
43 68
44 Number of sequences 69 This tool generates two outputs: a general summary and an optional gap stats file.
45 70
46 GC content in % 71 The general summary includes the following information:
47 72
48 If an optional genome size estimate is specified, then the NG50 length will also be calculated. 73 - Lengths: n50, min, max, median and average
74 - Number of base pairs: A, C, G, T, N, Total and Total_not_N
75 - Number of sequences
76 - GC content
49 77
50 ------ 78 In addition the optional gap stats BED file includes the information about gaps localization.
51 79 ]]>
52 Inputs:
53
54 Fasta dataset
55 </help> 80 </help>
56 <citations> 81 <citations>
57 <citation type="bibtex"> 82 <citation type="bibtex">
58 @UNPUBLISHED{Seemann_Gladman2012, 83 @UNPUBLISHED{Anmol_Kyran2021,
59 author = {Torsten Seemann and Simon Gladman}, 84 author = {Anmol Kyran},
60 title = {Fasta Statistics: Display summary statistics for a fasta file.}, 85 title = {Fasta Statistics: Display summary statistics for a fasta file.},
61 year = {2012}, 86 year = {2021},
62 url = {https://github.com/galaxyproject/tools-iuc}, 87 url = {https://github.com/galaxyproject/tools-iuc},
63 } 88 }
64 </citation> 89 </citation>
65 </citations> 90 </citations>
66 </tool> 91 </tool>