Merge pull request #174 from AlexandrovLab/ebv

Update v1.2.24: Fix ebv matrix generation
AlexandrovLab · Feb 9, 2024 · 518e22b · 518e22b
2 parents 9b12f7d + 70db87d
commit 518e22b
Show file tree

Hide file tree

Showing 3 changed files with 8 additions and 10 deletions.
diff --git a/SigProfilerMatrixGenerator/scripts/SigProfilerMatrixGeneratorFunc.py b/SigProfilerMatrixGenerator/scripts/SigProfilerMatrixGeneratorFunc.py
@@ -1417,11 +1417,12 @@ def SigProfilerMatrixGeneratorFunc(
             chrom_start = None
         if i != 1:
             for file in vcf_files:
-                chrom = file.split("_")[0]
+                if reference_genome.lower() != "ebv":
+                    chrom = file.split("_")[0]
+                else:
+                    chrom = "gi_82503188_ref_NC_007605"
                 if not os.path.exists(chrom_path + chrom + ".txt"):
                     continue
-                if reference_genome == "ebv":
-                    chrom = "_".join([x for x in file.split("_")[:-1]])
                 with open(vcf_path + file) as f:
                     lines = [line.strip().split() for line in f]
                 lines = sorted(lines, key=lambda x: (x[0], int(x[2])))
@@ -1953,7 +1954,7 @@ def SigProfilerMatrixGeneratorFunc(
             for file in vcf_files:
                 chrom = file.split("_")[0]
                 if reference_genome == "ebv":
-                    chrom = "_".join([x for x in file.split("_")[:-1]])
+                    chrom = "gi_82503188_ref_NC_007605"
                 with open(vcf_path + file) as f:
                     lines = [line.strip().split() for line in f]
                 lines = sorted(lines, key=lambda x: (x[0], int(x[2])))

diff --git a/SigProfilerMatrixGenerator/scripts/convert_input_to_simple_files.py b/SigProfilerMatrixGenerator/scripts/convert_input_to_simple_files.py
@@ -7,9 +7,6 @@
 from __future__ import print_function
 
 import os
-import pdb
-import re
-import sys
 
 from SigProfilerMatrixGenerator.scripts import MutationMatrixGenerator as spm
 
@@ -76,7 +73,7 @@ def convertVCF(project, vcf_path, genome, output_path, ncbi_chrom, log_file):
                         if len(line) == 0:
                             continue
                         chrom = line[0]
-                        if len(chrom) > 2:
+                        if len(chrom) > 2 and genome.lower() != "ebv":
                             chrom = chrom[3:]
                         if chrom in ncbi_chrom:
                             chrom = ncbi_chrom[chrom]

diff --git a/setup.py b/setup.py
@@ -3,7 +3,7 @@
 
 from setuptools import setup
 
-VERSION = "1.2.23"
+VERSION = "1.2.24"
 
 # remove the dist folder first if exists
 if os.path.exists("dist"):
@@ -23,7 +23,7 @@ def write_version_py(filename="SigProfilerMatrixGenerator/version.py"):
 # THIS FILE IS GENERATED FROM SIGPROFILEMATRIXGENERATOR SETUP.PY
 short_version = '%(version)s'
 version = '%(version)s'
-Update = 'v1.2.23: Add SV and CNV matrix generation to CLI. Add Dockerfile.'
+Update = 'v1.2.24: Fix ebv input processing.'
 
 	"""
     fh = open(filename, "w")