Mercurial > repos > fubar > egapx_runner

comparison nf/subworkflows/ncbi/shared/diamond/main.nf @ 0:d9c5c5b87fec draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/ncbi/egapx commit 8173d01b08d9a91c9ec5f6cb50af346edc8020c4
author fubar
date Sat, 03 Aug 2024 11:16:53 +0000
parents
children
comparison
equal deleted inserted replaced
-1:000000000000 0:d9c5c5b87fec
1 #!/usr/bin/env nextflow
2 nextflow.enable.dsl=2
3
4
5 /*
6 *Execution of:
7 * /netmnt/vast01/gpi/regr/GPIPE_REGR1/system/2024-03-27.prod.build25780/bin/diamond
8 * -asn-cache /netmnt/vast01/gpi/regr/GPIPE_REGR1/data00/Gavia_stellata/GP37025.85624/sequence_cache
9 * -blastp-args '--sam-query-len --comp-based-stats 0 --evalue 0.0001 --very-sensitive --max-hsps 3'
10 * -diamond-executable /netmnt/vast01/gpi/regr/GPIPE_REGR1/system/2024-03-27.prod.build25780/third-party/diamond/diamond
11 * -lds2 /netmnt/vast01/gpi/regr/GPIPE_REGR1/data00/Gavia_stellata/GP37025.85624/846757/prot_gnomon_prepare.8202002/out/LDS2
12 * -ofmt seq-align-set
13 * -output-dir /netmnt/vast01/gpi/regr/GPIPE_REGR1/data00/Gavia_stellata/GP37025.85624/846757/diamond.8202022/out
14 * -output-manifest /netmnt/vast01/gpi/regr/GPIPE_REGR1/data00/Gavia_stellata/GP37025.85624/846757/diamond.8202022/out/align.mft
15 * -output-prefix hits
16 * ## query is gnomon-made proteins 'gnl|GNOMON|23016146.p'
17 * ## query-fmt is <String, `fasta', `seq-ids'>
18 * -query-fmt seq-ids
19 * -query-manifest /netmnt/vast01/gpi/regr/GPIPE_REGR1/data00/Gavia_stellata/GP37025.85624/846757/diamond.8202022/inp/query_ids.mft
20 * ## subject is swiss-prot ids 'sp|A0A009IHW8.1|ABTIR_ACIB9'
21 * -subject-fmt seq-ids
22 * -subject-manifest /netmnt/vast01/gpi/regr/GPIPE_REGR1/data00/Gavia_stellata/GP37025.85624/846757/diamond.8202022/inp/subject_ids.mft
23 * -work-area /netmnt/vast01/gpi/regr/GPIPE_REGR1/data00/Gavia_stellata/GP37025.85624/846757/diamond.8202022/tmp
24
25 */
26
27 include {to_map; shellSplit } from '../../utilities'
28
29
30 swiss_prot_url='https://ftp.ncbi.nlm.nih.gov/genomes/TOOLS/EGAP/reference_sets/swissprot.asnb.gz'
31 process fetch_swiss_prot_asn {
32 input:
33 output:
34 path "output/swissprot.asnb", emit: "swiss_prot_asn"
35 script:
36 """
37 curl -O '$swiss_prot_url'
38 gunzip swissprot.asnb.gz
39 mkdir -p output
40 mv swissprot.asnb output/swissprot.asnb
41 """
42 stub:
43 """
44 mkdir -p output
45 touch output/swissprot.asnb
46 """
47 }
48
49 process get_swiss_prot_ids {
50 input:
51 path swiss_prot_asn
52 output:
53 path "output/swiss_prot_ids"
54 script:
55 """
56 mkdir -p output
57 lds2_indexer -db lds -source .
58 sqlite3 ./lds "SELECT txt_id FROM seq_id WHERE orig=1 AND int_id IS NULL;" > output/swiss_prot_ids
59 """
60 stub:
61 """
62 mkdir -p output
63 touch output/swiss_prot_ids
64 """
65 }
66
67 process run_diamond_egap {
68 input:
69 path gnomon_prot_ids
70 path swiss_prot_ids
71 path gnomon_prot_asn, stageAs: 'indexed/*'
72 path swiss_prot_asn, stageAs: 'indexed/*'
73 val params
74 output:
75 path "output/*"
76 script:
77 // print(params)
78 """
79
80 ###diamond_bin=`which diamond`
81 #diamond_egap uses GP_HOME to build paths to both some gp apps, and third-party
82 #GP_HOME needs to be the directory that contains third-party, and the directory that contains bin/<gp apps>
83 diamond_bin=\${GP_HOME}/third-party/diamond/diamond
84
85 mkdir -p ./asncache/
86
87 prime_cache -cache ./asncache/ -ifmt asnb-seq-entry -i ${gnomon_prot_asn} -oseq-ids /dev/null -split-sequences
88 prime_cache -cache ./asncache/ -ifmt asnb-seq-entry -i ${swiss_prot_asn} -oseq-ids /dev/null -split-sequences
89
90 mkdir ./output
91 mkdir ./work
92
93 echo ${params}
94 echo "${gnomon_prot_ids.join('\n')}" > query.mft
95 diamond_egap ${params} -asn-cache ./asncache/ -nogenbank -query-manifest query.mft -subject ${swiss_prot_ids} \
96 -output-dir ./output/ -work-area ./work/ -diamond-executable \${diamond_bin}
97 rm -rf ./work
98 """
99
100 stub:
101 """
102 mkdir -p output
103 touch output/diamond_output.asn
104 """
105 }
106
107