From ca7c85c719cb24d5dcc569c9841d47bb0c7b184d Mon Sep 17 00:00:00 2001
From: Fontrodona Nicolas <nicolas.fontrodona@ens-lyon.fr>
Date: Wed, 4 Nov 2020 10:37:17 +0100
Subject: [PATCH] src/visu/__main__.py src/visu/figure_maker.py: modification
to create a figure with two different list of exons
---
src/visu/__main__.py | 16 +--
src/visu/figure_maker.py | 206 +++++++++++++++++++++++++++------------
2 files changed, 150 insertions(+), 72 deletions(-)
diff --git a/src/visu/__main__.py b/src/visu/__main__.py
index b084670..bc53640 100644
--- a/src/visu/__main__.py
+++ b/src/visu/__main__.py
@@ -13,11 +13,11 @@ from pathlib import Path
from typing import List
-@lp.parse(design='file', region_bed='file',
+@lp.parse(design='file', region_beds='file',
nb_bin="nb_bin > 5", figure_type=['metagene', 'barplot'],
show_replicate=['y', 'n', 'Y', 'N'])
-def launcher(design: str, bw_folder: str, region_bed: str,
- region_name: str, nb_bin: int = 100,
+def launcher(design: str, bw_folder: str, region_beds: List[str],
+ region_names: List[str], nb_bin: int = 100,
figure_type: str = 'metagene', norm: str = 'None',
show_replicate: str = 'y', environment: List[int] = (0, 0),
border_names: List[str] = ('', ''),
@@ -31,8 +31,9 @@ def launcher(design: str, bw_folder: str, region_bed: str,
the last one contains the replicate of the condition.
:param bw_folder: The folder containing the bigwig file mentioned in \
the first column of the 'design' table.
- :param region_bed: A bed file containing the regions to visualise
- :param region_name: The name of the region analysed
+ :param region_beds: A list of bed files containing the regions to visualise
+ :param region_names: A list of names identifying regions insides \
+ the given beds.
:param nb_bin: The number of bins used to represents the regions of \
'region_bed'.
:param figure_type: The kind of representation wanted (barplot or metagene)
@@ -58,8 +59,9 @@ def launcher(design: str, bw_folder: str, region_bed: str,
norm = Path(norm)
if not norm.is_file():
raise FileNotFoundError(f"The file {norm} was not found")
- create_figure(Path(design), Path(bw_folder), Path(region_bed),
- region_name, nb_bin, figure_type, norm, show_rep,
+ reg_beds = [Path(p) for p in region_beds]
+ create_figure(Path(design), Path(bw_folder), reg_beds,
+ region_names, nb_bin, figure_type, norm, show_rep,
environment, border_names, Path(output))
diff --git a/src/visu/figure_maker.py b/src/visu/figure_maker.py
index 0f6615f..2da518f 100644
--- a/src/visu/figure_maker.py
+++ b/src/visu/figure_maker.py
@@ -7,7 +7,7 @@ Description:
"""
from pathlib import Path
-from typing import List, Union, Any
+from typing import List, Union, Any, Tuple
from doctest import testmod
from ..bed_handler.config import TestConfig
import pandas as pd
@@ -17,15 +17,16 @@ import matplotlib.pyplot as plt
from tqdm import tqdm
-def load_bed(bed: Path) -> List[List[Union[int, str]]]:
+def load_bed(bed: Path, bed_name: str) -> List[List[Union[int, str]]]:
"""
Read a bed file and return the lines within it.
:param bed: A bed file containing the regions of interest
+ :param bed_name: The name of the regions of interest inside the bed file
:return:The list of feature inside the bed
- >>> load_bed(TestConfig.gene_bed)[0]
- ['18', 28645943, 28682388, 1, 'DSC2', '-']
+ >>> load_bed(TestConfig.gene_bed, 'gene_test')[0]
+ ['18', 28645943, 28682388, 1, 'DSC2', '-', 'gene_test']
"""
list_regions = []
with bed.open('r') as inbed:
@@ -33,10 +34,34 @@ def load_bed(bed: Path) -> List[List[Union[int, str]]]:
if not line.startswith("#"):
cline = line.replace("\n", "").split("\t")
list_regions.append([cline[0], int(cline[1]), int(cline[2]),
- int(cline[3]), cline[4], cline[5]])
+ int(cline[3]), cline[4], cline[5],
+ bed_name])
return list_regions
+def load_beds(beds: List[Path], bed_names: List[str]
+ ) -> List[List[Union[int, str]]]:
+ """
+ Read a bed file and return the lines within it.
+
+ :param beds: A list of bed files containing the regions of interest
+ :param bed_names: A list of names indentifying regions insides the given \
+ beds.
+ :return:The list of feature inside the beds file
+
+ >>> load_beds([TestConfig.gene_bed, TestConfig.gene_bed],
+ ... ['gene1', 'gene2'])[0]
+ ['18', 28645943, 28682388, 1, 'DSC2', '-', 'gene1']
+ >>> load_beds([TestConfig.gene_bed, TestConfig.gene_bed],
+ ... ['gene1', 'gene2'])[-1]
+ ['13', 45967450, 45992516, 9, 'SLC25A30', '-', 'gene2']
+ """
+ regions = []
+ for i in range(len(beds)):
+ regions += load_bed(beds[i], bed_names[i])
+ return regions
+
+
def inspect_bigwig_regions(bw: Any, region: List,
replicate: str, nb_bin: int, resize: List[int],
condition_name: str,
@@ -54,30 +79,30 @@ def inspect_bigwig_regions(bw: Any, region: List,
:return: a table with the coverage of this region
>>> my_bw = pbw.open(str(TestConfig.small_bw))
- >>> region = ['1', 10, 25, 1, 'Test', '+']
- >>> inspect_bigwig_regions(my_bw, region, 'R1', 5, [4, 2], 'cond1')
- coverage bin condition replicate
- 0 0.000000 -2 cond1 R1
- 1 0.500000 -1 cond1 R1
- 2 75.000000 0 cond1 R1
- 3 20.000000 1 cond1 R1
- 4 10.000000 2 cond1 R1
- 5 4.666667 3 cond1 R1
- 6 2.000000 4 cond1 R1
- 7 1.000000 5 cond1 R1
- 8 0.500000 6 cond1 R1
- >>> region = ['1', 110, 133, 1, 'Test', '-']
- >>> inspect_bigwig_regions(my_bw, region, 'R1', 5, [4, 2], 'cond1')
- coverage bin condition replicate
- 0 0.00 -2 cond1 R1
- 1 12.50 -1 cond1 R1
- 2 42.00 0 cond1 R1
- 3 8.00 1 cond1 R1
- 4 4.25 2 cond1 R1
- 5 2.00 3 cond1 R1
- 6 2.00 4 cond1 R1
- 7 1.00 5 cond1 R1
- 8 1.00 6 cond1 R1
+ >>> mregion = ['1', 10, 25, 1, 'Test', '+', 'exon']
+ >>> inspect_bigwig_regions(my_bw, mregion, 'R1', 5, [4, 2], 'cond1')
+ coverage bin condition replicate region
+ 0 0.000000 -2 cond1 R1 exon
+ 1 0.500000 -1 cond1 R1 exon
+ 2 75.000000 0 cond1 R1 exon
+ 3 20.000000 1 cond1 R1 exon
+ 4 10.000000 2 cond1 R1 exon
+ 5 4.666667 3 cond1 R1 exon
+ 6 2.000000 4 cond1 R1 exon
+ 7 1.000000 5 cond1 R1 exon
+ 8 0.500000 6 cond1 R1 exon
+ >>> mregion = ['1', 110, 133, 1, 'Test', '-', 'exon2']
+ >>> inspect_bigwig_regions(my_bw, mregion, 'R1', 5, [4, 2], 'cond1')
+ coverage bin condition replicate region
+ 0 0.00 -2 cond1 R1 exon2
+ 1 12.50 -1 cond1 R1 exon2
+ 2 42.00 0 cond1 R1 exon2
+ 3 8.00 1 cond1 R1 exon2
+ 4 4.25 2 cond1 R1 exon2
+ 5 2.00 3 cond1 R1 exon2
+ 6 2.00 4 cond1 R1 exon2
+ 7 1.00 5 cond1 R1 exon2
+ 8 1.00 6 cond1 R1 exon2
"""
val = bw.stats(region[0], region[1], region[2], nBins=nb_bin, exact=True)
bins = list(range(len(val)))
@@ -110,6 +135,7 @@ def inspect_bigwig_regions(bw: Any, region: List,
df = pd.DataFrame(dic)
df['condition'] = [condition_name] * df.shape[0]
df['replicate'] = [replicate] * df.shape[0]
+ df['region'] = [region[6]] * df.shape[0]
return df
@@ -165,10 +191,31 @@ def create_full_table(df_exp: pd.DataFrame, regions: List[List],
return pd.concat(list_df, axis=0, ignore_index=True)
+def merge_condition_region_col(df: pd.DataFrame) -> Tuple[pd.DataFrame, str]:
+ """
+
+ :param df: A dataframe of mean coverage for each bin.
+ :return: The dataframe with the region of condition column merged and \
+ the name of the merged column
+ """
+ if len(df['region'].unique()) == 1:
+ condition_col = 'condition'
+ df.drop('region', axis=1, inplace=True)
+ elif len(df['condition'].unique()) == 1:
+ condition_col = 'region'
+ df.drop('condition', axis=1, inplace=True)
+ else:
+ condition_col = 'condition-region'
+ df[condition_col] = df['condition'] + "-" + df['region']
+ df.drop(['condition', 'region'], axis=1, inplace=True)
+ return df, condition_col
+
+
def create_df_summary(df_cov: pd.DataFrame, figure_type: str, nb_bin: int,
environment: List[int],
- region_name: str, order_condition: List[str]
- ) -> pd.DataFrame:
+ region_name: str, order_condition: List[str],
+ order_bed_name: List[str],
+ ) -> Tuple[pd.DataFrame, str]:
"""
summarize the data in df_cov.
@@ -180,21 +227,24 @@ def create_df_summary(df_cov: pd.DataFrame, figure_type: str, nb_bin: int,
the number of bin used to represent those surrounding regions.
:param region_name: the name of the region analysed
:param order_condition: The order of conditions
- :return: The summarised dataframe
+ :param order_bed_name: The order of bed name to respect
+ :return: The summarised dataframe and the condition col
"""
- df_sum = df_cov.groupby(['bin', 'condition', 'replicate']).mean() \
+ df_sum = df_cov.groupby(['bin', 'condition', 'region', 'replicate']) \
+ .mean() \
.reset_index()
if figure_type == "metagene":
- return df_sum
+ df_sum, condition_col = merge_condition_region_col(df_sum)
+ return df_sum, condition_col
if environment[0] != 0:
df_sum['location'] = df_cov['bin'].apply(
lambda x: f"before_{region_name}" if x < 0 else
f"after_{region_name}" if x >= nb_bin else region_name)
df_sum.drop('bin', axis=1, inplace=True)
if environment[0] != 0:
- col_merge = ['condition', 'replicate', 'location']
+ col_merge = ['condition', 'region', 'replicate', 'location']
else:
- col_merge = ['condition', 'replicate']
+ col_merge = ['condition', 'region', 'replicate']
df_sum = df_sum.groupby(col_merge).mean().reset_index()
if 'location' in df_sum.columns:
df_sum['location'] = pd.Categorical(
@@ -206,15 +256,21 @@ def create_df_summary(df_cov: pd.DataFrame, figure_type: str, nb_bin: int,
df_sum['condition'], ordered=True,
categories=order_condition
)
- df_sum.sort_values(['condition', 'location'], ascending=True,
+ df_sum['region'] = pd.Categorical(
+ df_sum['region'], ordered=True,
+ categories=order_bed_name
+ )
+ df_sum.sort_values(['condition', 'region', 'location'], ascending=True,
inplace=True)
- return df_sum
+ df_sum, condition_col = merge_condition_region_col(df_sum)
+ return df_sum, condition_col
def figure_metagene(df_sum: pd.DataFrame, show_replicate: bool,
border_names: List[str], nb_bin: int,
- environment: List[int], region_name: str,
- output: Path, norm: Union[int, Path]) -> None:
+ environment: List[int], bed_name: str,
+ output: Path, norm: Union[int, Path],
+ condition_col: str) -> None:
"""
Create a metagene figure on the region of interest.
@@ -226,18 +282,19 @@ def figure_metagene(df_sum: pd.DataFrame, show_replicate: bool,
nucleotide to represent around the region of interest and the second,
the number of bin used to represent those surrounding regions.
:param output: Folder where the figure will be created
- :param region_name: The region of interest
+ :param bed_name: The name of considered regions
:param norm: an integer corresponding to the bin used to normalise \
the samples or a file containing the normalisations to apply to \
each samples
+ :param condition_col: The name of the condition columns
"""
sns.set(context='poster', style='white')
if show_replicate:
- g = sns.relplot('bin', 'coverage', hue='condition', data=df_sum,
+ g = sns.relplot('bin', 'coverage', hue=condition_col, data=df_sum,
kind='line', style='replicate', ci=None,
height=12, aspect=1.7)
else:
- g = sns.relplot('bin', 'coverage', hue='condition', data=df_sum,
+ g = sns.relplot('bin', 'coverage', hue=condition_col, data=df_sum,
kind='line', ci="sd", height=12, aspect=1.7)
y_val = g.ax.get_ylim()[1] * 0.99
if border_names[0] != '':
@@ -250,11 +307,13 @@ def figure_metagene(df_sum: pd.DataFrame, show_replicate: bool,
g.set_xlabels('Bins')
g.set_ylabels('Coverage')
plt.subplots_adjust(top=0.9)
- title = f"Average coverage in region '{region_name}'"
+ tmp_bed_name = bed_name.replace("--", ", ")
+ title = f"Average coverage in region '{tmp_bed_name}'"
if environment[0] != 0:
title += f"\nand in their surrounding regions of {environment[0]} nt"
g.fig.suptitle(title)
- outfile_title = f"metagene_{region_name}_{nb_bin}bin_" \
+ tmp_bed_name = bed_name.replace("--", "-")
+ outfile_title = f"metagene_{tmp_bed_name}_{nb_bin}bin_" \
f"{environment[0]}_nt-around-{environment[1]}-bin"
if isinstance(norm, int):
outfile_title += f"_b{norm}_norm"
@@ -268,7 +327,8 @@ def figure_metagene(df_sum: pd.DataFrame, show_replicate: bool,
def figure_barplot(df_sum: pd.DataFrame, show_replicate: bool,
nb_bin: int,
environment: List[int], region_name: str,
- output: Path, norm: Union[int, Path]) -> None:
+ output: Path, norm: Union[int, Path],
+ condition_col: str) -> None:
"""
Create a barplot figure on the region of interest.
@@ -283,22 +343,25 @@ def figure_barplot(df_sum: pd.DataFrame, show_replicate: bool,
:param norm: an integer corresponding to the bin used to normalise \
the samples or a file containing the normalisations to apply to \
each samples
+ :param condition_col: The name of the condition columns
"""
sns.set(context='poster', style='white')
if show_replicate:
- g = sns.catplot(x="condition", y="coverage", hue="replicate",
+ g = sns.catplot(x=condition_col, y="coverage", hue="replicate",
kind="bar", data=df_sum, height=12, aspect=1.77,
ci=None)
else:
- g = sns.catplot(x="condition", y="coverage",
+ g = sns.catplot(x=condition_col, y="coverage",
kind="bar", data=df_sum, height=12, aspect=1.77,
ci='sd')
g.set_xlabels('')
g.set_ylabels('Coverage')
plt.subplots_adjust(top=0.9)
- title = f"Average coverage in region '{region_name}'"
+ rgt = region_name.replace('--', ', ')
+ title = f"Average coverage in region '{rgt}'"
g.fig.suptitle(title)
- outfile_title = f"barplot_{region_name}_{nb_bin}bin_" \
+ rgt = region_name.replace('--', '-')
+ outfile_title = f"barplot_{rgt}_{nb_bin}bin_" \
f"{environment[0]}_nt-around-{environment[1]}-bin"
if isinstance(norm, int):
outfile_title += f"_b{norm}_norm"
@@ -326,8 +389,8 @@ def bin_normalisation(df: pd.DataFrame, norm: Union[int, Path],
raise ValueError(f"the bin {norm} was not found in the coverage "
f"dataframe.")
df_val = df.loc[df['bin'] == norm,
- ['coverage', 'condition', 'replicate']] \
- .groupby(['condition', 'replicate']).mean().reset_index()
+ ['coverage', 'condition', 'region', 'replicate']] \
+ .groupby(['condition', 'region', 'replicate']).mean().reset_index()
df_val.rename({"coverage": "coef"}, axis=1, inplace=True)
noutfile = outfile.parent / 'coef_table' / \
(outfile.name.replace(".txt.gz", "") +
@@ -336,14 +399,19 @@ def bin_normalisation(df: pd.DataFrame, norm: Union[int, Path],
df_val.to_csv(noutfile, sep="\t", index=False)
else:
df_val = pd.read_csv(norm, sep="\t")
- df = df.merge(df_val, how="left", on=['condition', 'replicate'])
+ if len(df_val['region'].unique()) > 1:
+ df = df.merge(df_val, how="left", on=['condition', 'region',
+ 'replicate'])
+ else:
+ df_val.drop('region', axis=1, inplace=True)
+ df = df.merge(df_val, how="left", on=['condition', 'replicate'])
df['coverage'] = df['coverage'] / df['coef']
df.drop('coef', axis=1, inplace=True)
return df
-def create_figure(design: Path, bw_folder: Path, region_bed: Path,
- region_name: str, nb_bin: int = 100,
+def create_figure(design: Path, bw_folder: Path, region_beds: List[Path],
+ bed_names: List[str], nb_bin: int = 100,
figure_type: str = 'metagene',
norm: Union[int, Path, None] = None,
show_replicate: bool = True, environment: List[int] = (0, 0),
@@ -358,8 +426,9 @@ def create_figure(design: Path, bw_folder: Path, region_bed: Path,
the last one contains the replicate of the condition.
:param bw_folder: The folder containing the bigwig file mentioned in \
the first column of the 'design' table.
- :param region_bed: A bed file containing the regions to visualise
- :param region_name: The name of the region analysed
+ :param region_beds: A list of bed files containing the regions to visualise
+ :param bed_names: A list of names identifying regions insides the given \
+ beds.
:param nb_bin: The number of bins used to represents the regions of \
'region_bed'.
:param figure_type: The kind of representation wanted (barplot or metagene)
@@ -374,10 +443,15 @@ def create_figure(design: Path, bw_folder: Path, region_bed: Path,
:param border_names: The name of the borders
:param output: Folder where the results will be created
"""
+ if len(region_beds) != len(bed_names):
+ raise IndexError("Parameter region_beds and bed_names should "
+ "have the same length")
df_exp = pd.read_csv(design, sep="\t")
- regions = load_bed(region_bed)
- region_bed_name = region_bed.name.replace('.bed', '')
- outfile = f'tmp_cov_table_{design.name}_{region_bed_name}_{nb_bin}bin_' \
+ regions = load_beds(region_beds, bed_names)
+ region_bed_name = "-".join([b.name.replace('.bed', '')
+ for b in region_beds])
+ outfile = f'tmp_cov_table_{design.name.replace(".txt", "")}' \
+ f'_{region_bed_name}_{nb_bin}bin_' \
f'{environment[0]}_nt-around-{environment[1]}-bin'
if isinstance(norm, int):
outfile += f'_bin{norm}_norm'
@@ -397,20 +471,22 @@ def create_figure(design: Path, bw_folder: Path, region_bed: Path,
for condition in df_exp['condition'].to_list():
if condition not in ordered_condition:
ordered_condition.append(condition)
- df_sum = create_df_summary(df_cov, figure_type, nb_bin, environment,
- region_name, ordered_condition)
+ region_kind = "--".join(bed_names)
+ df_sum, cond_col = create_df_summary(df_cov, figure_type, nb_bin,
+ environment, region_kind,
+ ordered_condition, bed_names)
if figure_type == "metagene":
figure_metagene(df_sum, show_replicate, border_names, nb_bin,
- environment, region_name, output, norm)
+ environment, region_kind, output, norm, cond_col)
else:
if 'location' in df_sum.columns:
for cur_region in df_sum['location'].unique():
df_tmp = df_sum.loc[df_sum['location'] == cur_region, :]
figure_barplot(df_tmp, show_replicate, nb_bin, environment,
- cur_region, output, norm)
+ cur_region, output, norm, cond_col)
else:
figure_barplot(df_sum, show_replicate, nb_bin, environment,
- region_name, output, norm)
+ region_kind, output, norm, cond_col)
if __name__ == "__main__":
--
GitLab