From 49c0594a4d2d3e5a16d9fef72d60b39228d51fb5 Mon Sep 17 00:00:00 2001
From: Fontrodona Nicolas <nicolas.fontrodona@ens-lyon.fr>
Date: Mon, 15 Mar 2021 18:18:09 +0100
Subject: [PATCH] src/gc_content/gc_stats.py: fix mann_whitney and make_stat
function used when more than two lists of exons are used
---
src/gc_content/gc_stats.py | 16 +++++++++-------
1 file changed, 9 insertions(+), 7 deletions(-)
diff --git a/src/gc_content/gc_stats.py b/src/gc_content/gc_stats.py
index 778022b..483460c 100644
--- a/src/gc_content/gc_stats.py
+++ b/src/gc_content/gc_stats.py
@@ -21,10 +21,12 @@ def mann_whitney(df: pd.DataFrame) -> List[Any]:
:return: The name of the groups and their p-value
"""
regions = df['region'].unique()
+ res = []
for r1, r2 in combinations(regions, 2):
v1 = df.loc[df['region'] == r1, "gc_content"]
v2 = df.loc[df['region'] == r2, "gc_content"]
- return [r1, r2, mannwhitneyu(v1, v2)[-1]]
+ res.append([r1, r2, mannwhitneyu(v1, v2)[-1]])
+ return res
def make_stat(df: pd.DataFrame) -> List[List[Any]]:
@@ -35,12 +37,12 @@ def make_stat(df: pd.DataFrame) -> List[List[Any]]:
"""
if len(df['location'].unique()) == 1:
return [mann_whitney(df)]
- else:
- list_pval = []
- for loc in df['location'].unique():
- df_tmp = df.loc[df["location"] == loc, :]
- res = mann_whitney(df_tmp)
+ list_pval = []
+ for loc in df['location'].unique():
+ df_tmp = df.loc[df["location"] == loc, :]
+ result = mann_whitney(df_tmp)
+ for res in result:
res[0] = (loc, res[0])
res[1] = (loc, res[1])
list_pval.append(res)
- return list_pval
+ return list_pval
--
GitLab