fix: add fake protein/nucleotide to the inputs to BUSCO proteome/tran…

…scriptome runs in the absense any models being present in the fasta

minos/__init__.py

@@ -7,3 +7,9 @@
 __copyright__ = "Copyright 2019-2023 Earlham Institute"
 __version__ = pkg_resources.require("minos")[0].version
 
+DEFAULT_FAKE_PROT = pkg_resources.resource_filename(
+    "minos.etc", "fake_prot.fasta"
+)
+DEFAULT_FAKE_NUC = pkg_resources.resource_filename(
+    "minos.etc", "fake_nuc.fasta"
+)

minos/etc/fake_nuc.fasta

@@ -0,0 +1,2 @@
+>fake_nuc
+TACAATACTATCCCGCGCGGTTGCGGATATCCCAGTCTACAGCGTTGAGTGAGACACATG

minos/etc/fake_prot.fasta

@@ -0,0 +1,2 @@
+>fake_prot
+FKEPTN

minos/scripts/busco_splitter.py

@@ -1,3 +1,14 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+"""
+Script to split and check fasta for BUSCO analysis
+"""
+
+import os
+from minos import (
+    DEFAULT_FAKE_PROT,
+    DEFAULT_FAKE_NUC,
+)
 def split_fasta(fastafile, split_files):
     out = None
     for line in open(fastafile):
@@ -18,3 +29,23 @@ def split_fasta(fastafile, split_files):
     for f in split_files.values():
         f.close()
 
+def copy_file(_in, label, _out):
+    with open(_in, 'r') as fh:
+        fhdata = fh.read()
+
+    fhdata = fhdata.replace('>', f'>{label}_')
+
+    with open(_out, 'w') as fh:
+        fh.write(fhdata)
+
+
+def check_split_fasta(format, fasta_files):
+    for label, path in fasta_files.items():
+        fasta = path.name
+        if os.path.exists(fasta) and os.stat(fasta).st_size == 0:
+            if format == "prot":
+                print(f"INFO: Copy fake protein '{DEFAULT_FAKE_PROT}' to empty protein fasta file '{fasta}' to prevent BUSCO analysis failure")
+                copy_file(DEFAULT_FAKE_PROT, label, fasta)
+            if format == "nuc":
+                print(f"INFO: Copy fake nucleotide '{DEFAULT_FAKE_NUC}' to empty nucleotide fasta file '{fasta}' to prevent BUSCO analysis failure")
+                copy_file(DEFAULT_FAKE_NUC, label, fasta)

minos/zzz/minos_run.smk

@@ -950,9 +950,10 @@ rule split_proteins_prepare:
 	log:
 		os.path.join(BUSCO_PATH, "logs", "split_proteins_prepare.log")
 	run:
-		from minos.scripts.busco_splitter import split_fasta
+		from minos.scripts.busco_splitter import split_fasta, check_split_fasta
 		fasta_files = {tm: open(os.path.join(BUSCO_PATH, "runs", "proteins_prepare", "input", tm + ".proteins.fasta"), "w") for tm in config["data"]["transcript_models"]}
 		split_fasta(input[0], fasta_files)
+		check_split_fasta("prot", fasta_files)
 
 rule busco_proteins_prepare:
 	input:
@@ -1013,9 +1014,10 @@ rule split_transcripts_prepare:
 	log:
 		os.path.join(BUSCO_PATH, "logs", "split_transcripts_prepare.log")
 	run:
-		from minos.scripts.busco_splitter import split_fasta
+		from minos.scripts.busco_splitter import split_fasta, check_split_fasta
 		fasta_files = {tm: open(os.path.join(BUSCO_PATH, "runs", "transcripts_prepare", "input", tm + ".cdna.fasta"), "w") for tm in config["data"]["transcript_models"]}
 		split_fasta(input[0], fasta_files)
+		check_split_fasta("nuc", fasta_files)
 
 rule busco_transcripts_prepare:
 	input: