New pftools module for PROSITE and HAMAP

README.md

@@ -317,7 +317,7 @@ nextflow run ebi-pf-team/interproscan6 \
   -r 6.0.0 \
   -profile docker \
   -c licensed.conf \
-  --input /path/to/sequences.faayour.fasta \
+  --input /path/to/sequences.faa \
   --applications deeptmhmm,phobius,signalp_euk,signalp_prok \
   --use-gpu
 ```

modules/prosite/patterns/main.nf → modules/pftools/main.nf

@@ -1,11 +1,13 @@
+import groovy.io.FileType
 import groovy.json.JsonOutput
 import groovy.json.JsonSlurper
 import uk.ac.ebi.interpro.Location
 import uk.ac.ebi.interpro.Match
 import uk.ac.ebi.interpro.Signature
 import uk.ac.ebi.interpro.SignatureLibraryRelease
 
-process RUN_PFSCAN {
+
+process RUN_PSSCAN {
     /*
     The ps_scan.pl script is a wrapper for the pfscan tool that is provided by the
     pftools developers. It automates running pfscan for all provided patterns and
@@ -34,20 +36,20 @@ process RUN_PFSCAN {
 }
 
 
-process PARSE_PFSCAN {
+process PARSE_PSSCAN {
     label    'mem_low', 'time_veryshort'
     executor 'local'
 
     input:
-        tuple val(meta), val(pfscan_out)
+        tuple val(meta), val(ps_scan_out)
 
     output:
-        tuple val(meta), path("prositepatterns.json")
+        tuple val(meta), path("ps_scan.json")
 
     exec:
     Map<String, Map<String, Match>> patternsMatches = [:]
     SignatureLibraryRelease library = new SignatureLibraryRelease("PROSITE patterns", null)
-    pfscan_out.eachLine { line ->
+    ps_scan_out.eachLine { line ->
         line = line.trim()
         if (!line || line.startsWith("pfscanV3 is not meant to be used with a single profile")) {
             return
@@ -79,7 +81,75 @@ process PARSE_PFSCAN {
         matchObj.addLocation(location)
     }
 
-    def outputFilePath = task.workDir.resolve("prositepatterns.json")
+    def outputFilePath = task.workDir.resolve("ps_scan.json")
     def json = JsonOutput.toJson(patternsMatches)
     new File(outputFilePath.toString()).write(json)
 }
+
+process RUN_PFSEARCH {
+    label 'mem_min', 'time_medium', 'dynamic', 'ips6_container'
+
+    input:
+    tuple val(meta), path(fasta)
+    path dirpath
+    val profiles_dir
+
+    output:
+    tuple val(meta), stdout
+
+    script:
+    """
+    find ${dirpath}/${profiles_dir} -type f | while read profile; do
+        pfsearchV3 -f -o 7 -t ${task.cpus} "\${profile}" "${fasta}"
+    done
+    """
+}
+
+process PARSE_PFSEARCH {
+    label    'mem_low', 'time_veryshort'
+    executor 'local'
+
+    input:
+    tuple val(meta), val(pfsearch_out)
+    val signature_library
+    val dirpath
+    val blacklist_file
+
+    output:
+    tuple val(meta), path("pfsearch.json")
+
+    exec:
+    Map matches = [:]
+    SignatureLibraryRelease library = new SignatureLibraryRelease(signature_library, null)
+    def toSkip = []
+    if (dirpath && blacklist_file) {
+        toSkip = new File("${dirpath.toString()}/${blacklist_file}").readLines()
+    }
+
+    pfsearch_out.eachLine { line ->
+        def fields = line.split()
+        assert fields.size() == 10
+        String modelAccession = fields[0].split("\\|")[0]
+        if (toSkip && (modelAccession in toSkip)) {
+            return // skip flagged accessions
+        }
+
+        String seqId = fields[3]
+        int start = fields[4].toInteger()
+        int end = fields[5].toInteger()
+        Double score = Double.parseDouble(fields[7])
+        String alignment = fields[9]
+        String cigarAlignment = Match.encodeCigarAlignment(alignment)
+
+        matches.computeIfAbsent(seqId) { [:] }
+        Match matchObj = matches[seqId].computeIfAbsent(modelAccession) {
+            new Match(modelAccession, new Signature(modelAccession, library))
+        }
+        Location location = new Location(start, end, score, alignment, cigarAlignment)
+        matchObj.addLocation(location)
+    }
+    def outputFilePath = task.workDir.resolve("pfsearch.json")
+    def json = JsonOutput.toJson(matches)
+    new File(outputFilePath.toString()).write(json)
+}
+

subworkflows/hamap/main.nf

@@ -1,4 +1,4 @@
-include { RUN_PFSEARCH; PARSE_PFSEARCH } from  "../../modules/prosite/profiles"
+include { RUN_PFSEARCH ; PARSE_PFSEARCH } from  "../../modules/pftools"
 
 workflow HAMAP {
     take:

subworkflows/prosite/patterns/main.nf

@@ -1,4 +1,4 @@
-include { RUN_PFSCAN; PARSE_PFSCAN } from  "../../../modules/prosite/patterns"
+include { RUN_PSSCAN; PARSE_PSSCAN } from  "../../../modules/pftools"
 
 workflow PROSITE_PATTERNS {
     take:
@@ -8,14 +8,14 @@ workflow PROSITE_PATTERNS {
     evafile         // str repr of the path to the eva file in the data dir -> datadir/evafile
 
     main:
-    RUN_PFSCAN(
+    RUN_PSSCAN(
         ch_seqs,
         dirpath,
         datfile,
         evafile
     )
 
-    ch_prosite = PARSE_PFSCAN(RUN_PFSCAN.out)
+    ch_prosite = PARSE_PSSCAN(RUN_PSSCAN.out)
 
     emit:
     ch_prosite

subworkflows/prosite/profiles/main.nf

@@ -1,4 +1,4 @@
-include { RUN_PFSEARCH ; PARSE_PFSEARCH } from  "../../../modules/prosite/profiles"
+include { RUN_PFSEARCH ; PARSE_PFSEARCH } from  "../../../modules/pftools"
 
 workflow PROSITE_PROFILES {
     take:

tests/unit_tests/modules/scan/hamap/parse_hamap.nf.test

@@ -1,7 +1,7 @@
 nextflow_process {
 
     name "Test Process PARSE_PFSEARCH"
-    script "modules/prosite/profiles/main.nf"
+    script "modules/pftools/main.nf"
     process "PARSE_PFSEARCH"
 
     test("Should run without failures") {

tests/unit_tests/modules/scan/prosite/patterns/main.nf.test

@@ -1,8 +1,8 @@
 nextflow_process {
 
-    name "Test Process PARSE_PFSCAN"
-    script "modules/prosite/patterns/main.nf"
-    process "PARSE_PFSCAN"
+    name "Test Process PARSE_PSSCAN"
+    script "modules/pftools/main.nf"
+    process "PARSE_PSSCAN"
 
     test("Should run without failures") {
 

tests/unit_tests/modules/scan/prosite/profiles/main.nf.test

@@ -1,7 +1,7 @@
 nextflow_process {
 
     name "Test Process PARSE_PFSEARCH"
-    script "modules/prosite/profiles/main.nf"
+    script "modules/pftools/main.nf"
     process "PARSE_PFSEARCH"
 
     test("Should run without failures") {