Merge pull request #10 from Finn-Lab/implement_only_ips_gff_support

Add support for protein FASTA input

README.md

@@ -33,15 +33,22 @@ sanntis test/files/BGC0001472.fna
 conda deactivate sanntis
 ```
 
-#### Use precomputed InterProScan outputs
+#### Support of preprocessed InterProScan outputs
 
-SanntiS can be executed using preprocessed InterProScan outputs along with a GenBank (GBK) file specifying the coding sequences (CDSs). This integration facilitates a streamlined analysis pipeline for bioinformatics applications, allowing for enhanced functionality and user flexibility.
+SanntiS can be executed using preprocessed InterProScan outputs along with a GenBank (GBK) file specifying the coding sequences (CDSs). This integration increases user flexibility.
 ```bash
 conda activate sanntis
 sanntis --ip-file test/files/BGC0001472.fna.prodigal.faa.gff3 test/files/BGC0001472.fna.prodigal.faa.gb
 conda deactivate sanntis
 ```
 
+Additionally, the --ip-file option can be provided with a protein FASTA file containing headers formatted according to Prodigal's convention. In this case, the --is_protein flag must be included to indicate that the sequence file is a proteins FASTA. 
+```bash
+conda activate sanntis
+sanntis --is_protein --ip-file test/files/BGC0001472.fna.prodigal.faa.gff3 test/files/BGC0001472.fna.prodigal.faa
+conda deactivate sanntis
+```
+
 ###  Docker:
 
 #### Get InterProsScan data:

sanntis/_cli.py

@@ -35,9 +35,18 @@ def main(args=None):
     parser.add_argument(
         "seq_file",
         type=str,
-        help="input nucleotide sequence file. FASTA or GBK. mandatory",
+        help=(
+            "Input sequence file. Supported formats: nucleotide FASTA, GBK, or protein FASTA. "
+            "If the file is a protein FASTA, it must use Prodigal output headers and must be accompanied "
+            "by the --is_protein flag. Mandatory."
+        ),
         metavar="SEQUENCE_FILE",
     )
+    parser.add_argument(
+        "--is_protein",
+        action="store_true",
+        help="Specify if the input SEQUENCE_FILE is a protein FASTA file. Will only process sequences with headers formatted like Prodigal protein outputs.",
+    )
     parser.add_argument(
         "-v",
         "--version",
@@ -66,62 +75,62 @@ def main(args=None):
         dest="score",
         default=None,
         type=float,
-        help="validation filter threshold. overrides --greed",
+        help="Validation filter threshold. overrides --greed",
         metavar="FLOAT",
     )
     parser.add_argument(
         "--meta",
         dest="meta",
         default="True",
         type=str,
-        help="prodigal option meta [default True]",
+        help="Prodigal option meta [default True]",
         metavar="True|False",
     )
     parser.add_argument(
         "--outdir",
         default=os.getcwd(),
         dest="outdir",
         type=str,
-        help="output directory [default $PWD/SEQUENCE_FILE.sanntis]",
+        help="Output directory [default $PWD/SEQUENCE_FILE.sanntis]",
         metavar="DIRECTORY",
     )
     parser.add_argument(
         "--outfile",
         dest="outfile",
         type=str,
-        help="output file [default outdir/SEQUENCE_FILE.sanntis.gff]",
+        help="Output file [default outdir/SEQUENCE_FILE.sanntis.gff]",
         metavar="FILE",
     )
     parser.add_argument(
         "--minimal",
         dest="minimal_out",
         default="True",
         type=str,
-        help="minimal output in a gff3 file [default True]",
+        help="Minimal output in a gff3 file [default True]",
         metavar="True|False",
     )
     parser.add_argument(
         "--antismash_output",
         dest="antismash_out",
         default="False",
         type=str,
-        help="write results in antiSMASH 6.0 JSON specification output [default False]",
+        help="Write results in antiSMASH 6.0 JSON specification output [default False]",
         metavar="True|False",
     )
     parser.add_argument(
         "--refined",
         dest="ref_b",
         default="False",
         type=str,
-        help="annotate high probability borders [default False]",
+        help="Annotate high probability borders [default False]",
         metavar="True|False",
     )
     parser.add_argument(
         "--cpu",
         dest="cpu",
         default=1,
         type=int,
-        help="cpus for INTERPROSCAN and HMMSCAN",
+        help="Cpus for INTERPROSCAN and HMMSCAN",
         metavar="INT",
     )
 
@@ -153,6 +162,7 @@ def main(args=None):
     log.info("preprocessing files")
     preprocess = Preprocess(
             os.path.abspath(args.seq_file), 
+            args.is_protein,
             args.ip_file, 
             args.meta, 
             args.cpu, 

sanntis/modules/BGCdetection.py

@@ -18,6 +18,7 @@
 import pickle
 from itertools import groupby
 
+from Bio import SeqIO
 import numpy as np
 
 os.environ['TF_CPP_MIN_LOG_LEVEL'] = '1'
@@ -93,48 +94,43 @@ def transformEmeraldHmm(self, hmmFile):
                 spl = l.split()
                 self.entriesDct.setdefault(spl[3], []).append(spl[0])
 
-    def transformCDSpredToCDScontigs(self, cdsPredFile, f):
+    def transformCDSpredToCDScontigs(self, cdsPredFile, file_format):
 
         if not os.path.isfile(cdsPredFile):
             log.exception(f"{cdsPredFile} file not found")
+
+        if file_format == "fasta":
+
+            _prodigal_pattern = re.compile(
+                r"_\d+\s#\s(\d+)\s#\s(\d+)\s#\s(-?1)\s#\sID=(\d+_\d+);partial=(\d{2});start_type="
+                r"(\w+);rbs_motif=(.+);rbs_spacer=(\S+);gc_cont=(\d+\.\d+)"
+            )
+
+            for record in SeqIO.parse(open(cdsPredFile, "r"), file_format):
+                header = record.description
+                prodigal_match = _prodigal_pattern.search(header)
+                if not prodigal_match:
+                    log.warning(
+                        f"Protein {record.id} does not follow the Prodigal header format. "
+                    )
+                    continue
+                start = int(prodigal_match.group(1))
+                end = int(prodigal_match.group(2))
+                protein_id = record.id
 
-        with open(cdsPredFile, "r") as h:
-
-            if f == "fasta":
-
-                for l in h:
-
-                    if l[0] != ">":
-                        continue
-
-                    spl = l.split()
-                    start, end = int(spl[2]), int(spl[4])
-
-                    self.contigsDct.setdefault(
-                        "_".join(spl[0].split("_")[:-1])[1:], []
-                    ).append((spl[0][1:].strip(), (start, end)))
-
-            elif f == "genbank":
-
-                from Bio import SeqIO
-
-                recs = list(SeqIO.parse(open(cdsPredFile, "r"), "gb"))
-
-                for rec in recs:
-                    for f in rec.features:
-                        if f.type == "CDS":
-
+                self.contigsDct.setdefault(
+                    "_".join(record.id.split("_")[:-1]), []
+                ).append((record.id, (start, end)))
 
-                            start, end = int(f.location.start) + 1, int(f.location.end)
+        elif file_format == "genbank":
+
+            for record in SeqIO.parse(open(cdsPredFile, "r"), file_format):
+                for f in record.features:
+                    if f.type == "CDS":
+                        start, end = int(f.location.start) + 1, int(f.location.end)
+                        protein_id = f.qualifiers["protein_id"][0].strip().replace(' ','') if "protein_id" in f.qualifiers else f.qualifiers["locus_tag"][0].strip().replace(' ','') 
 
-                            self.contigsDct.setdefault(rec.id, []).append(
-                                (
-                                    f.qualifiers["protein_id"][0].strip().replace(' ','') # replace to avoid long id bug in gb files
-                                    if "protein_id" in f.qualifiers
-                                    else f.qualifiers["locus_tag"][0].strip().replace(' ',''),
-                                    (start, end),
-                                )
-                            )
+                        self.contigsDct.setdefault(record.id, []).append( (protein_id,(start, end)))
 
     def buildMatrices(self):
 
@@ -161,7 +157,6 @@ def buildMatrices(self):
                 ]
                 self.annDct[contig][ix][modiAnn] = 1
 
-
     def predictAnn(self, colapseFunc=max):
 
         log.info("Predict BGC probability w/ TensorFlow")

sanntis/modules/Preproc.py

@@ -25,9 +25,10 @@
 class Preprocess:
     """External tools needed for sanntis bgc detection"""
 
-    def __init__(self, seq_file, ip_file, meta, cpus, outdir):
+    def __init__(self, seq_file, seqfile_is_proteins, ip_file, meta, cpus, outdir):
 
         self.seq_file = seq_file
+        self.seqfile_is_proteins = seqfile_is_proteins
         self.ip_file = ip_file
         self.meta = meta
         self.cpus = int(cpus)
@@ -130,8 +131,9 @@ def process_sequence(self):
         """ CDS prediction on sequence file"""
 
         self.check_fmt()
+
         if self.fmt == "fasta":
-            self.outFaa = self.runProdigal()
+            self.outFaa = self.seq_file if self.seqfile_is_proteins else self.runProdigal()
         elif self.fmt == "genbank":
             self.outFaa = self.gbkToProdigal()
         else:

sanntis/pkg_info.json

@@ -1,6 +1,6 @@
 {
     "name": "sanntis",
-    "version": "0.9.3.5",
+    "version": "0.9.4.0",
     "description": "SMBGC detection tool",
     "author": "Santiago Sanchez Fragoso",
     "author_email": "[email protected]",