improve mosek license check, make reduce_fasta a library

rename pre_filtered bed to unfiltered_gain and put in cnvkit directory

AmpliconSuite-pipeline.py

@@ -14,7 +14,7 @@
 import tarfile
 import time
 
-from paalib import check_reference, cnv_prefilter
+from paalib import check_reference, cnv_prefilter, reduce_fasta
 from paalib._version import __ampliconsuitepipeline_version__
 
 
@@ -166,9 +166,7 @@ def run_cnvkit(ckpy_path, nthreads, outdir, bamfile, seg_meth='cbs', normal=None
     logging.info("Running CNVKit batch\n")
     if normal and not args.ref == "GRCh38_viral":
         # create a version of the stripped reference
-        scripts_dir = os.path.dirname(os.path.abspath(__file__)) + "/scripts/"
-        strip_cmd = "python {}reduce_fasta.py -r {} -c {} -o {}".format(scripts_dir, ref_fasta, ref_genome_size_file, outdir)
-        call(strip_cmd, shell=True)
+        reduce_fasta.reduce_fasta(ref_fasta, ref_genome_size_file, outdir)
         base = os.path.basename(ref_fasta) # args.ref is the name, ref is the fasta
         stripRefG = outdir + os.path.splitext(base)[0] + "_reduced" + "".join(os.path.splitext(base)[1:])
         logging.debug("Stripped reference: " + stripRefG)
@@ -824,11 +822,16 @@ def contains_spaces(file_path):
 
     if args.run_AA:
         if not os.path.exists(os.environ["HOME"] + "/mosek/mosek.lic") and not "MOSEKLM_LICENSE_FILE" in os.environ:
-            logging.error("--run_AA set, but MOSEK license not found!")
+            logging.error("--run_AA set, but MOSEK license not found in $HOME/mosek/")
             sys.exit(1)
 
-        elif "MOSEKLM_LICENSE_FILE" in os.environ and not os.path.exists(os.environ["MOSEKLM_LICENSE_FILE"] + "/mosek.lic"):
-                logging.error("--run_AA set, but MOSEK license not found!")
+        elif "MOSEKLM_LICENSE_FILE" in os.environ:
+            if os.environ["MOSEKLM_LICENSE_FILE"].endswith("mosek.lic"):
+                logging.error("MOSEKLM_LICENSE_FILE should be the path of the directory of the license, not the full path. Please update your .bashrc, and run 'source ~/.bashrc'")
+                sys.exit(1)
+
+            elif not os.path.exists(os.environ["MOSEKLM_LICENSE_FILE"] + "/mosek.lic"):
+                logging.error("--run_AA set, but MOSEK license not found in " + os.environ["MOSEKLM_LICENSE_FILE"])
                 sys.exit(1)
 
     runCNV = None
@@ -975,6 +978,7 @@ def contains_spaces(file_path):
         centromere_dict = get_ref_centromeres(args.ref)
         chr_sizes = get_ref_sizes(ref_genome_size_file)
         # coordinate CNV calling
+        cnvkit_output_directory = None
         if runCNV == "CNVkit":
             cnvkit_output_directory = args.output_directory + sname + "_cnvkit_output/"
             if not os.path.exists(cnvkit_output_directory):
@@ -1001,17 +1005,19 @@ def contains_spaces(file_path):
         sample_info_dict["sample_cnv_bed"] = args.cnv_bed
 
         if not args.no_filter and not args.cnv_bed.endswith("_AA_CNV_SEEDS.bed"):
-            if not args.cnv_bed.endswith("_CNV_CALLS_pre_filtered.bed"):
+            if not args.cnv_bed.endswith("_CNV_CALLS_pre_filtered.bed") and not args.cnv_bed.endswith("_CNV_CALLS_unfiltered_gains.bed"):
+                pfilt_odir = cnvkit_output_directory if cnvkit_output_directory else args.output_directory
                 args.cnv_bed = cnv_prefilter.prefilter_bed(args.cnv_bed, args.ref, centromere_dict, chr_sizes,
-                                                           args.cngain, args.output_directory)
+                                                           args.cngain, pfilt_odir)
 
             amplified_interval_bed = run_amplified_intervals(args.aa_python_interpreter, args.cnv_bed, args.bam,
                                                              outdir, sname, args.cngain, args.cnsize_min)
 
         elif args.no_filter and runCNV:
-            if not args.cnv_bed.endswith("_CNV_CALLS_pre_filtered.bed"):
+            if not args.cnv_bed.endswith("_CNV_CALLS_pre_filtered.bed") and not args.cnv_bed.endswith("_CNV_CALLS_unfiltered_gains.bed"):
+                pfilt_odir = cnvkit_output_directory if cnvkit_output_directory else args.output_directory
                 args.cnv_bed = cnv_prefilter.prefilter_bed(args.cnv_bed, args.ref, centromere_dict, chr_sizes,
-                                                           args.cngain, args.output_directory)
+                                                           args.cngain, pfilt_odir)
                 logging.info("Skipping amplified_intervals.py step due to --no_filter")
 
         else:

README.md

@@ -38,8 +38,9 @@ AmpliconSuite-pipeline can also be run through Nextflow, using the [nf-core/circ
 
 ### Option B: Conda or Mamba
 ```bash 
-conda create -n ampsuite && conda activate ampsuite
-conda install -c bioconda -c mosek -c conda-forge ampliconsuite mosek
+conda create -n ampsuite python=3.10 && conda activate ampsuite
+conda install -c bioconda -c conda-forge ampliconsuite 
+conda install -c mosek mosek
 
 # then run the installer script to finalize the locations of the data repo and mosek license 
 wget https://raw.githubusercontent.com/AmpliconSuite/AmpliconSuite-pipeline/master/install.sh

docker/run_paa_docker.py

@@ -149,6 +149,9 @@ def metadata_helper(metadata_args):
 try:
     # deprecated install of mosek license
     MOSEKLM_LICENSE_FILE = os.environ['MOSEKLM_LICENSE_FILE']
+    if MOSEKLM_LICENSE_FILE.endswith("mosek.lic"):
+        sys.stderr.write(
+            "MOSEKLM_LICENSE_FILE should be the path of the directory of the license, not the full path. Please update your .bashrc, and run 'source ~/.bashrc'\n")
     if not os.path.exists(MOSEKLM_LICENSE_FILE + "/mosek.lic"):
         raise KeyError
 

paalib/_version.py

@@ -1 +1 @@
-__ampliconsuitepipeline_version__ = "1.3.1"
+__ampliconsuitepipeline_version__ = "1.3.2"

paalib/cnv_prefilter.py

@@ -197,7 +197,7 @@ def prefilter_bed(bedfile, ref, centromere_dict, chr_sizes, cngain, outdir):
 
     merged_filt_ivald = merge_intervals(filt_ivald, cn_cut=cngain, require_same_cn=True, ref=ref)
     final_filt_entries = ivald_to_ilist(merged_filt_ivald)
-    bname = outdir + "/" + bedfile.rsplit("/")[-1].rsplit(".bed")[0] + "_pre_filtered.bed"
+    bname = outdir + "/" + bedfile.rsplit("/")[-1].rsplit(".bed")[0] + "_unfiltered_gains.bed"
     with open(bname, 'w') as outfile:
         for entry in final_filt_entries:
             outfile.write("\t".join([str(x) for x in entry]) + "\n")

scripts/reduce_fasta.py → paalib/reduce_fasta.py

@@ -39,6 +39,18 @@ def fasta_reader(fasta_file, chroms_to_get):
     return fasta_dict
 
 
+def reduce_fasta(ref, chrom, outname=""):
+    chrList = getRelChrs(chrom)
+    seqD = fasta_reader(ref, chrList)
+    base = os.path.basename(ref)
+    refGOutName = os.path.splitext(base)[0] + "_reduced" + "".join(os.path.splitext(base)[1:])
+    print("Writing stripped FASTA\n")
+    with open(outname + refGOutName, 'w') as outfile:
+        for i in chrList:
+            outfile.write(">" + i + "\n")
+            outfile.write(seqD[i] + "\n")
+
+
 if __name__ == '__main__':
     # Parses the command line arguments
     parser = argparse.ArgumentParser(description="Reduce FASTA to entries in list")
@@ -50,14 +62,5 @@ def fasta_reader(fasta_file, chroms_to_get):
     if not args.outname:
         args.outname = ""
 
-    chrList = getRelChrs(args.chrom)
-    seqD = fasta_reader(args.ref, chrList)
-    base = os.path.basename(args.ref)
-    refGOutName = os.path.splitext(base)[0] + "_reduced" + "".join(os.path.splitext(base)[1:])
-    print("Writing stripped FASTA\n")
-    with open(args.outname + refGOutName, 'w') as outfile:
-        for i in chrList:
-            outfile.write(">" + i + "\n")
-            outfile.write(seqD[i] + "\n")
-
+    reduce_fasta(args.ref, args.chrom, args.outname)
     sys.exit()

singularity/run_paa_singularity.py

@@ -162,6 +162,7 @@ def metadata_helper(metadata_args):
 
 
 if not os.path.exists(os.environ['HOME'] + "/mosek/mosek.lic"):
+    sys.stdout.write("Python detected $HOME was set to: " + os.environ['HOME'])
     sys.stderr.write(
         "Mosek license (mosek.lic) file not found in $HOME/mosek/. Please see README for instructions.\n")
     sys.exit(1)