From f0a7a47df628428d328c890c61c680a63db87c3a Mon Sep 17 00:00:00 2001
From: Eric Normandeau <eric.normandeau.qc@gmail.com>
Date: Mon, 29 May 2023 14:06:10 -0400
Subject: [PATCH] Improve options

---
 01_scripts/08_replace_coordinates_in_vcf.py | 10 +++++-----
 02_infos/snplift_config.sh                  | 15 +++++++++------
 2 files changed, 14 insertions(+), 11 deletions(-)

diff --git a/01_scripts/08_replace_coordinates_in_vcf.py b/01_scripts/08_replace_coordinates_in_vcf.py
index 19507fc..5a88b8b 100755
--- a/01_scripts/08_replace_coordinates_in_vcf.py
+++ b/01_scripts/08_replace_coordinates_in_vcf.py
@@ -39,11 +39,11 @@ def reverse_complement(seq):
     input_vcf = sys.argv[1]
     input_correspondence = sys.argv[2]
     unique_pos = sys.argv[3]
-    correct_id = sys.argv[4]
+    correct_id = int(sys.argv[4])
     id_column = sys.argv[5]
-    correct_alleles = sys.argv[6]
+    correct_alleles = int(sys.argv[6])
     allele_columns = sys.argv[7]
-    sort_output = sys.argv[8]
+    sort_output = int(sys.argv[8])
     output_vcf = sys.argv[9]
 except:
     print(__doc__)
@@ -95,11 +95,11 @@ def reverse_complement(seq):
 
             # Skip already treated positions
             if unique_pos:
-                if l[2] in viewed_positions:
+                if (l[0], l[1]) in viewed_positions:
                     continue
 
                 else:
-                    viewed_positions.add(l[2])
+                    viewed_positions.add((l[0], l[1]))
 
             # Sorting
             if sort_output:
diff --git a/02_infos/snplift_config.sh b/02_infos/snplift_config.sh
index c857269..9d78c33 100644
--- a/02_infos/snplift_config.sh
+++ b/02_infos/snplift_config.sh
@@ -21,7 +21,7 @@ export SKIP_VISUALIZATION=1 # Avoid creating a plot to explore features. These a
                             #   for debugging [0, 1].
 
 # Number of CPUs
-export NCPUS=10             # Number of cores to use (around 10 and maximum 20 is recommended)
+export NCPUS=20             # Number of cores to use (around 10 and maximum 20 is recommended)
                             #   For less than 100K SNPs, 1 to 4 cores is a good choice.
                             #   For less than 1M SNPs, 10 cores is a good choice.
                             #   Above this, 20 cores is going to be slightly faster.
@@ -33,13 +33,16 @@ export WINDOW_LENGTH=300    # Sise of sequences kept on both sides of each SNP.
 
 # Number of neighbours to use
 export NUM_NEIGHBOURS=20    # Number of neighbour SNPs to consider when trying to salvage
-                            #   more dubious SNPs using local correlations of positions
+                            #   more dubious SNPs using local correlations of positions.
 
 # Do final corrections to VCF file
-export UNIQUE_POS=1         # Keep only the first locus encountered for each position [0, 1].
 export CORRECT_ID=1         # Recompute the ID column from columns 1 and 2 [0, 1].
 export ID_COLUMN=3          # ID column number [int]. Only used for `CORRECT_ID=1`.
 export CORRECT_ALLELES=1    # Reverse complement alleles when new genome region is inversed [0, 1].
-export ALLELE_COLUMNS=4,5   # Columns of alleles to reverse complement. Only used for `CORRECT_ALLELES=1`
-export SORT_OUTPUT=1        # Sort VCF or output file according to the two first columns
-                            #   WARNING: For big VCFs, this will require a lot of RAM
+export ALLELE_COLUMNS=4,5   # Columns of alleles to reverse complement. Only used for `CORRECT_ALLELES=1`.
+
+# Memory hungry options
+export UNIQUE_POS=1         # Keep only the first locus encountered for each position [0, 1]. Memory hungry.
+export SORT_OUTPUT=1        # Sort VCF or output file according to the two first columns.
+                            #   WARNING: For big VCFs, this will require a lot of RAM, around 3X the size
+                            #   of the output file.