From aee24d7d0d288f7623533031e8c1bd3b15448c2c Mon Sep 17 00:00:00 2001
From: Fontrodona Nicolas <nicolas.fontrodona@ens-lyon.fr>
Date: Tue, 3 Nov 2020 15:21:03 +0100
Subject: [PATCH] src/bed_handler/select_regulated_near_ctcf_exons.py: add a
parameter include0 to include exons containing CTCF if wanted. Add a
parameter near_ctcf that allows to recover regulated exons far from a CTCF
site
---
.../select_regulated_near_ctcf_exons.py | 94 ++++++++++++++-----
1 file changed, 71 insertions(+), 23 deletions(-)
diff --git a/src/bed_handler/select_regulated_near_ctcf_exons.py b/src/bed_handler/select_regulated_near_ctcf_exons.py
index 4389c9c..6894170 100644
--- a/src/bed_handler/select_regulated_near_ctcf_exons.py
+++ b/src/bed_handler/select_regulated_near_ctcf_exons.py
@@ -11,10 +11,11 @@ from .get_gene_regulated_by_ddx import load_sipp_vs_ctcf, format_exon_bed
import pandas as pd
from doctest import testmod
from .filter_gene import filter_bed
+import warnings
def filter_ctcf_distance_table(df: pd.DataFrame, reg: str, threshold: int,
- location: str) -> pd.DataFrame:
+ location: str, include0: bool) -> pd.DataFrame:
"""
Filter the dataframe to recover only regulated exons near CTCF.
@@ -22,57 +23,74 @@ def filter_ctcf_distance_table(df: pd.DataFrame, reg: str, threshold: int,
:param reg: The regulation by CTCF
:param threshold: The threshold distance
:param location: The location of interest
+ :param include0: True to include exons containing a CTCF site if the \
+ threshold is greater than 0.
:return: The filtered dataframe
>>> cdf = load_sipp_vs_ctcf(TestConfig.sipp_vs_ctcf,
... format_exon_bed(TestConfig.exon_bed, TestConfig.gene_bed))
- >>> rdf = filter_ctcf_distance_table(cdf, 'down', 4, 'upstream')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'down', 4, 'upstream', False)
>>> rdf[['exon_name', 'dist', 'group', 'id']]
exon_name dist group id
0 DSC2_1 -4 siPP_DOWN 1_1
1 DSC2_2 -3 siPP_DOWN 1_2
2 DSC2_3 -2 siPP_DOWN 1_3
- >>> rdf = filter_ctcf_distance_table(cdf, 'down', 2, 'upstream')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'down', 4, 'upstream', True)
+ >>> rdf[['exon_name', 'dist', 'group', 'id']]
+ exon_name dist group id
+ 0 DSC2_1 -4 siPP_DOWN 1_1
+ 1 DSC2_2 -3 siPP_DOWN 1_2
+ 2 DSC2_3 -2 siPP_DOWN 1_3
+ 4 DSC2_5 0 siPP_DOWN 1_5
+ >>> rdf = filter_ctcf_distance_table(cdf, 'down', 2, 'upstream', False)
>>> rdf[['exon_name', 'dist', 'group', 'id']]
exon_name dist group id
2 DSC2_3 -2 siPP_DOWN 1_3
- >>> rdf = filter_ctcf_distance_table(cdf, 'down', 2, 'downstream')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'down', 2, 'downstream', False)
>>> rdf[['exon_name', 'dist', 'group', 'id']]
exon_name dist group id
5 DSC2_6 1 siPP_DOWN 1_6
6 DSC2_7 2 siPP_DOWN 1_7
- >>> rdf = filter_ctcf_distance_table(cdf, 'down', 2, 'both')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'down', 2, 'both', False)
>>> rdf[['exon_name', 'dist', 'group', 'id']]
exon_name dist group id
2 DSC2_3 -2 siPP_DOWN 1_3
5 DSC2_6 1 siPP_DOWN 1_6
6 DSC2_7 2 siPP_DOWN 1_7
- >>> rdf = filter_ctcf_distance_table(cdf, 'up', 2, 'both')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'up', 2, 'both', False)
>>> rdf[['exon_name', 'dist', 'group', 'id']]
exon_name dist group id
3 DSC2_4 -1 siPP_UP 1_4
- >>> rdf = filter_ctcf_distance_table(cdf, 'up', 2, 'downstream')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'up', 2, 'downstream', False)
>>> rdf[['exon_name', 'dist', 'group', 'id']]
Empty DataFrame
Columns: [exon_name, dist, group, id]
Index: []
- >>> rdf = filter_ctcf_distance_table(cdf, 'all', 2, 'both')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'all', 2, 'both', False)
>>> rdf[['exon_name', 'dist', 'group', 'id']]
exon_name dist group id
2 DSC2_3 -2 siPP_DOWN 1_3
3 DSC2_4 -1 siPP_UP 1_4
5 DSC2_6 1 siPP_DOWN 1_6
6 DSC2_7 2 siPP_DOWN 1_7
- >>> filter_ctcf_distance_table(cdf, 'lul', 2, 'both')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'all', 2, 'both', True)
+ >>> rdf[['exon_name', 'dist', 'group', 'id']]
+ exon_name dist group id
+ 2 DSC2_3 -2 siPP_DOWN 1_3
+ 3 DSC2_4 -1 siPP_UP 1_4
+ 4 DSC2_5 0 siPP_DOWN 1_5
+ 5 DSC2_6 1 siPP_DOWN 1_6
+ 6 DSC2_7 2 siPP_DOWN 1_7
+ >>> filter_ctcf_distance_table(cdf, 'lul', 2, 'both', False)
Traceback (most recent call last):
...
ValueError: reg parameter should be one in: ['down', 'up', 'all']
- >>> filter_ctcf_distance_table(cdf, 'up', 2, 'xd')
+ >>> filter_ctcf_distance_table(cdf, 'up', 2, 'xd', False)
Traceback (most recent call last):
...
ValueError: location parameter should be in ['upstream', \
'downstream', 'both']
- >>> rdf = filter_ctcf_distance_table(cdf, 'down', 0, 'both')
+ >>> rdf = filter_ctcf_distance_table(cdf, 'down', 0, 'both', False)
>>> rdf[['exon_name', 'dist', 'group', 'id']]
exon_name dist group id
4 DSC2_5 0 siPP_DOWN 1_5
@@ -87,45 +105,75 @@ def filter_ctcf_distance_table(df: pd.DataFrame, reg: str, threshold: int,
df = df.loc[df["group"] == f"siPP_{reg.upper()}", :]
if location == "upstream":
df = df.loc[(df["dist"] >= threshold * -1) &
- (df["dist"] < 0), :]
+ (df["dist"] < 0 if not include0
+ else df["dist"] <= 0), :]
elif location == "downstream":
df = df.loc[(df["dist"] <= threshold) &
- (df["dist"] > 0), :]
+ (df["dist"] > 0 if not include0
+ else df["dist"] >= 0), :]
else:
if threshold == 0:
df = df.loc[abs(df["dist"]) == 0, :]
else:
- df = df.loc[(abs(df["dist"]) <= threshold) &
- (df["dist"] != 0), :]
+ if not include0:
+ df = df.loc[(abs(df["dist"]) <= threshold) &
+ (df["dist"] != 0), :]
+ else:
+ df = df.loc[abs(df["dist"]) <= threshold, :]
return df
def create_bed_ctcf_exon(reg: str, threshold: int,
- location: str) -> None:
+ location: str, include0: bool = False,
+ near_ctcf: bool = True) -> None:
"""
Filter the dataframe to recover only regulated exons near CTCF.
:param reg: The regulation by CTCF
:param threshold: The threshold distance
:param location: The location of interest
+ :param include0: True to include exons containing a CTCF site if the \
+ threshold is greater than 0.
+ :param near_ctcf: True to recover exons near CTCF False to recover \
+ those far from CTCF
"""
if threshold < 0:
threshold = 0
if threshold == 0:
location = "both"
- df = load_sipp_vs_ctcf(BedConfig.sipp_vs_ctcf,
- format_exon_bed(BedConfig.exon_bed,
- BedConfig.gene_bed))
- df = filter_ctcf_distance_table(df, reg, threshold, location)
- list_exons = df['id'].to_list()
+ if not include0 and threshold == 0:
+ warnings.warn("include0 must be True when threshold = 0."
+ "Setting include0 to true !")
+ include0 = True
+ if not near_ctcf and not include0:
+ warnings.warn("include0 must be True when near_ctcf is False")
+ include0 = True
+ i0 = "with0" if include0 else "without0"
+ df_reg = load_sipp_vs_ctcf(BedConfig.sipp_vs_ctcf,
+ format_exon_bed(BedConfig.exon_bed,
+ BedConfig.gene_bed))
+ df = filter_ctcf_distance_table(df_reg, reg, threshold, location, include0)
+ if near_ctcf:
+ name_near = ""
+ list_exons = df['id'].to_list()
+ else:
+ name_near = "Far_"
+ if reg != "all":
+ tmp_exons = df_reg.loc[df_reg["group"] == f"siPP_{reg.upper()}",
+ "id"].to_list()
+ else:
+ tmp_exons = df_reg['id'].to_list()
+ list_exons = [e for e in tmp_exons if e not in df['id'].to_list()]
list_genes = [int(exon.split('_')[0]) for exon in list_exons]
df_exon = filter_bed(BedConfig.exon_bed, list_exons)
df_gene = filter_bed(BedConfig.gene_bed, list_genes)
df_exon.to_csv(BedConfig.bed.output /
- f"CTCF_{threshold}_{location}_ddx_{reg}_exon.bed", sep="\t",
+ f"{name_near}CTCF_{threshold}_{location}_ddx_{reg}_{i0}_exon.bed",
+ sep="\t",
index=False)
df_gene.to_csv(BedConfig.bed.output /
- f"CTCF_{threshold}_{location}_ddx_{reg}_gene.bed", sep="\t",
+ f"{name_near}CTCF_{threshold}_{location}_ddx_{reg}_{i0}_gene.bed",
+ sep="\t",
index=False)
--
GitLab