From 9784bc1729dc0cd76594a1a38f92972785e30e97 Mon Sep 17 00:00:00 2001
From: Tanushree Tunstall <tanu@tunstall.in>
Date: Wed, 24 Feb 2021 09:56:36 +0000
Subject: [PATCH] updated count.py with indel and stop codon count

---
 scripts/count.py                        | 30 ++++++++++++++++++++++---
 scripts/plotting/other_plots_combined.R | 14 +-----------
 2 files changed, 28 insertions(+), 16 deletions(-)

diff --git a/scripts/count.py b/scripts/count.py
index 444af46..9121cda 100755
--- a/scripts/count.py
+++ b/scripts/count.py
@@ -276,13 +276,37 @@ foo_count = foo.loc[foo[dr_muts_col].str.contains('del', na = False, regex = Tru
 
 bar = meta_data[other_muts_col].value_counts()
 bar = bar.reset_index(name = 'values')
-bar.columns = [other_muts_col, 'dr_muts_count'] #64
+bar.columns = [other_muts_col, 'other_muts_count'] #64
 bar_count = bar.loc[bar[other_muts_col].str.contains('del', na = False, regex = True, case = False)]
 
 
 tot = len(foo_count) + len(bar_count)
 n_del = tot/len(meta_data)
-n_del*100
+n_del*100 #0.6
 
+#============================
+foo2 = meta_data[dr_muts_col].value_counts()
+foo2 = foo2.reset_index(name = 'values')
+foo2.columns = [dr_muts_col, 'dr_muts_count'] #171
+foo2_count = foo2.loc[foo2[dr_muts_col].str.contains('ins', na = False, regex = True, case = False) ]
 
-baz = meta_data.loc[meta_data[dr_muts_col].str.contains(nssnp_match, na = False, regex = True, case = False) | meta_data[other_muts_col].str.contains(nssnp_match, na = False, regex = True, case = False) ]
+bar2 = meta_data[other_muts_col].value_counts()
+bar2 = bar2.reset_index(name = 'values')
+bar2.columns = [other_muts_col, 'pther_muts_count'] #64
+bar2_count = bar2.loc[bar2[other_muts_col].str.contains('ins', na = False, regex = True, case = False)]
+
+tot2 = len(foo2_count) + len(bar2_count)
+n_ins = tot2/len(meta_data)
+n_ins*100 #0.5
+
+#=========
+del_ins_tot = n_del*100 + n_ins*100
+#==========
+# stop codons
+baz_count = foo.loc[foo[dr_muts_col].str.contains('\*', na = False, regex = True, case = False) ]
+baz_count2 = bar.loc[bar[other_muts_col].str.contains('\*', na = False, regex = True, case = False)]
+tot3 = len(baz_count) + len(baz_count2)
+n_stop = tot3/len(meta_data)
+n_stop*100 #0.11
+
+all_tot = n_del*100 + n_ins*100 + n_stop*100 #1.2
diff --git a/scripts/plotting/other_plots_combined.R b/scripts/plotting/other_plots_combined.R
index ca0632b..d927808 100644
--- a/scripts/plotting/other_plots_combined.R
+++ b/scripts/plotting/other_plots_combined.R
@@ -131,7 +131,6 @@ p2 = ggplot(df_lf_foldx, aes(x = mutation_info
 
 p2 
 
-
 ##########################################################################
 #==============
 # Plot 3: LIG
@@ -248,21 +247,10 @@ print(Outplot_labelled)
 
 dev.off()
 
-
 #---------
 # plot 3: legend
 #---------
 #svg(plot_point_legend, width = 6, height = 7)
 #OutPlot3 =  legend
 #print(OutPlot3)
-#dev.off()
-
-
-
-
-
-
-
-
-
-
+#dev.off()
\ No newline at end of file