Mercurial > repos > rliterman > csp2

comparison CSP2/subworkflows/snpdiffs/main.nf @ 0:01431fa12065

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload"
author rliterman
date Mon, 02 Dec 2024 10:40:55 -0500
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:01431fa12065
1 // Screening and SNP Pipeline processing
2 output_directory = file(params.output_directory)
3 log_directory = file(params.log_directory)
4 screen_log_dir = file(params.screen_log_dir)
5 snp_log_dir = file(params.snp_log_dir)
6 snp_directory = file(params.snp_directory)
7
8 if(params.tmp_dir == ""){
9 temp_dir = ""
10 } else{
11 temp_dir = file(params.temp_dir)
12 }
13 ref_id_file = file(params.ref_id_file)
14
15 ref_mode = params.ref_mode
16
17 // Assess whether to rescue edge-filtered SNPs
18 edge_rescue = "${params.rescue}" == "norescue" ? "norescue" : "rescue"
19
20 // Set paths for output files
21 all_snpdiffs_list = file("${log_directory}/All_SNPDiffs.txt")
22 snp_dirs_list = file("${log_directory}/SNP_Dirs.txt")
23 screening_results_file = file("${output_directory}/Screening_Results.tsv")
24 isolate_data_file = file("${output_directory}/Isolate_Data.tsv")
25 snpdiffs_summary_file = file("${output_directory}/Raw_MUMmer_Summary.tsv")
26
27 // Get QC thresholds
28 min_cov = params.min_cov.toFloat()
29 min_length = params.min_len.toInteger()
30 min_iden = params.min_iden.toFloat()
31 reference_edge = params.ref_edge.toInteger()
32 query_edge = params.query_edge.toInteger()
33 max_missing = params.max_missing.toFloat()
34 n_ref = params.n_ref.toInteger()
35
36 workflow {
37 main:
38 // Run SNP pipeline
39 runSNPPipeline(query_data: all_snpdiffs, reference_data: ref_id_file)
40 }
41
42 workflow runScreen {
43
44 take:
45 all_snpdiffs
46
47 main:
48
49 all_snpdiffs
50 .unique{it -> it[2]}
51 .collect()
52 | screenSNPDiffs
53 }
54
55 process screenSNPDiffs{
56
57 input:
58 val(all_snpdiffs)
59
60 script:
61
62 screenDiffs = file("${projectDir}/bin/screenSNPDiffs.py")
63 """
64 $params.load_python_module
65 $params.load_bedtools_module
66 python $screenDiffs --snpdiffs_file "${all_snpdiffs_list}" --log_dir "${screen_log_dir}" --min_cov "${min_cov}" --min_len "${min_length}" --min_iden "${min_iden}" --ref_edge "${reference_edge}" --query_edge "${query_edge}" --density_windows "${params.dwin}" --max_snps "${params.wsnps}" --trim_name "${params.trim_name}" --output_file "${screening_results_file}" --ref_id "${ref_id_file}" --tmp_dir "${temp_dir}"
67 """
68 }
69
70 workflow runSNPPipeline{
71 take:
72 all_snpdiffs
73 reference_data
74
75 main:
76
77 query_snpdiffs = all_snpdiffs.map{tuple(it[0],it[2])}
78 ref_snpdiffs = all_snpdiffs.map{tuple(it[1],it[2])}
79
80 stacked_snpdiffs = query_snpdiffs.concat(ref_snpdiffs)
81 .collect().flatten().collate(2)
82
83 snp_dirs = stacked_snpdiffs
84 .combine(reference_data)
85 .filter{it -> it[0].toString() == it[2].toString()}
86 .map{it -> tuple(it[0],it[1])}
87 .groupTuple(by:0)
88 .map { ref, diff_files -> tuple( ref.toString(), diff_files.collect() ) }
89 | runSnpPipeline
90
91 //snp_dirs.collect() | compileResults
92 }
93
94 process compileResults{
95
96 executor = 'local'
97 cpus = 1
98 maxForks = 1
99
100 input:
101 val(snp_directories)
102
103 script:
104
105 compile_script = file("${projectDir}/bin/compileSNPResults.py")
106 snp_dirs_list.write(snp_directories.join("\n")+ "\n")
107 """
108 $params.load_python_module
109 python $compile_script --snp_dirs_file "${snp_dirs_list}" --output_directory "${snp_directory}" --isolate_data_file "${isolate_data_file}" --mummer_data_file "${snpdiffs_summary_file}"
110 """
111 }
112
113 process runSnpPipeline{
114
115 input:
116 tuple val(reference_id),val(diff_files)
117
118 output:
119 stdout
120
121 script:
122
123 snp_script = file("${projectDir}/bin/runSNPPipeline.py")
124
125 // Set + create output directory
126 snp_dir = file("${snp_directory}/${reference_id}")
127 snp_dir.mkdirs()
128
129 // Write SNPDiffs list
130 out_snpdiffs = file("${snp_dir}/SNPDiffs.txt")
131 out_snpdiffs.write(diff_files.join("\n")+ "\n")
132 """
133 $params.load_python_module
134 $params.load_bedtools_module
135 python $snp_script --reference_id "${reference_id}" --output_directory "${snp_dir}" --snpdiffs_file "${out_snpdiffs}" --log_directory "${snp_log_dir}" --min_cov "${min_cov}" --min_len "${min_length}" --min_iden "${min_iden}" --ref_edge "${reference_edge}" --query_edge "${query_edge}" --density_windows "${params.dwin}" --max_snps "${params.wsnps}" --trim_name "${params.trim_name}" --max_missing "${max_missing}" --tmp_dir "${temp_dir}" --rescue "${edge_rescue}"
136 echo -n $snp_dir
137 """
138 }