From f05e0585a97d431272eaf2282423304268249068 Mon Sep 17 00:00:00 2001
From: Fontrodona Nicolas <nicolas.fontrodona@ens-lyon.fr>
Date: Fri, 20 Nov 2020 16:12:10 +0100
Subject: [PATCH] src/find_interaction_cluster/*.py: replace list parameters
weights and global_weights by int paramet weight and global_weight
---
src/find_interaction_cluster/__main__.py | 24 +++---
.../community_finder.py | 75 ++++++-------------
.../create_ppi_files.py | 50 ++++++-------
.../nt_and_community.py | 42 +++++------
.../sf_and_communities.py | 37 +++++----
5 files changed, 95 insertions(+), 133 deletions(-)
diff --git a/src/find_interaction_cluster/__main__.py b/src/find_interaction_cluster/__main__.py
index 7fecb019..a196eeb3 100644
--- a/src/find_interaction_cluster/__main__.py
+++ b/src/find_interaction_cluster/__main__.py
@@ -20,12 +20,13 @@ from ..logging_conf import logging_def
@lp.parse
-def launcher(weight: List[int] = (1),
- global_weight: List[int] = (0),
+def launcher(weight: int = 1,
+ global_weight: int = 0,
same_gene: bool = True,
ps: int = ConfigGraph.cpu,
html_fig: bool = False, feature: str = 'exon', region: str = '',
iteration_ppi: int = 1000,
+ project: str = "GSM1018963_GSM1018964",
logging_level: str = "DISABLE"):
"""
Script used to find communities inside exon co-localized within a project
@@ -41,9 +42,11 @@ def launcher(weight: List[int] = (1),
:param html_fig: True to display the html figure (default False).
:param feature: The feature we want to analyse (default 'exon')
:param region: The region of a gene to analyse (used only if feature \
- is 'gene') (default '').
+ is 'gene') (default '') (can be 'gene', 'exon', 'intron').
:param logging_level: The level of data to display (default 'DISABLE')
:param iteration_ppi: the number of iteration for ppi analysis
+ :param project: The project name of interest \
+ (only used is global_weight = 0).
:param ps: The number of processes to use
"""
logging_def(ConfigGraph.community_folder, __file__, logging_level)
@@ -53,15 +56,14 @@ def launcher(weight: List[int] = (1),
same_gene = True
if not ConfigGraph.get_hipmcl_prog().is_file():
install_hipmcl("INFO")
- multiple_community_launcher(1, weight, global_weight, same_gene, html_fig,
- feature, logging_level)
- multiple_stat_launcher(ps, weight, global_weight, same_gene, feature,
- logging_level)
- multiple_nt_lmm_launcher(ps, weight, global_weight, same_gene, feature,
- region,
- logging_level)
+ multiple_community_launcher(weight, global_weight, project, same_gene,
+ html_fig, feature, logging_level)
+ multiple_stat_launcher(ps, weight, global_weight, project, same_gene,
+ feature, logging_level)
+ multiple_nt_lmm_launcher(ps, weight, global_weight, project,
+ same_gene, feature, region, logging_level)
if feature == "gene":
- ppi_stat_launcher(ps, weight, global_weight, same_gene,
+ ppi_stat_launcher(weight, global_weight, project, same_gene,
ConfigGraph.ppi_threshold, iteration_ppi,
logging_level)
diff --git a/src/find_interaction_cluster/community_finder.py b/src/find_interaction_cluster/community_finder.py
index b4180848..d65d5a81 100644
--- a/src/find_interaction_cluster/community_finder.py
+++ b/src/find_interaction_cluster/community_finder.py
@@ -253,8 +253,8 @@ def get_figure_title(project, weight, global_weight, same_gene, feature):
:param feature: The kind of analysed features
:return: A figure title
"""
- title = f"Co-localisation between {feature}s having a weight greater than " \
- f"{weight} in "
+ title = f"Co-localisation between {feature}s having a weight greater " \
+ f"than {weight} in "
if global_weight == 0:
title += f"the project {project}"
else:
@@ -347,7 +347,7 @@ def community_finder(weight: int, global_weight: int, project: str = "",
logging.debug('Done !')
-def get_projects(global_weight: int) -> List[str]:
+def get_projects(global_weight: int, project: str) -> str:
"""
Get projects name.
@@ -355,69 +355,38 @@ def get_projects(global_weight: int) -> List[str]:
the global weight is equal to 0 then then density figure are calculated \
by project, else all projet are merge together and the interaction \
seen in `global_weight` project are taken into account
- :return: The list of the project to consider
+ :param project: The name of a project
+ :return: The project to consider
"""
if global_weight != 0:
- return [f'Global-weight-{global_weight}']
+ return f'Global-weight-{global_weight}'
else:
- return ConfigGraph.good_projects
+ return project
-def get_projects_name(global_weights: List[int]) -> Tuple[List[str], Dict]:
- """
- Get projects name given a list of global_weight and a dictionary linking,
- each project name to it's corresponding global weight.
-
- :param global_weights: The list of global weights to consider. if \
- the global weight is equal to 0 then then density figure are calculated \
- by project, else all projet are merge together and the interaction \
- seen in `global_weight` project are taken into account
- :return: project names and a dictionary linking,
- each name to it's corresponding global weight.
- """
- dic = {}
- projects = []
- for global_weight in global_weights:
- tmp = get_projects(global_weight)
- projects += tmp
- for p in tmp:
- dic[p] = global_weight
- return projects, dic
-
-
-def multiple_community_launcher(ps: int, weights: List[int],
- global_weights: List[int],
- same_gene: bool, html_fig: bool = False,
+def multiple_community_launcher(weight: int,
+ global_weight: int,
+ project: str,
+ same_gene: bool,
+ html_fig: bool = False,
feature: str = 'exon',
logging_level: str = "DISABLE"):
"""
- :param ps: The number of processes we want to use.
- :param weights: The list of weights of interaction to consider
- :param global_weights: The list global weights to consider. if \
- the global weight is equal to 0 then then density figure are calculated \
- by project, else all projcet are merge together and the interaction \
- seen in `global_weight` project are taken into account
+ :param weight: The weight of interaction to consider
+ :param global_weight: The global weighs to consider. if \
+ the global weight is equal to 0 then the project `project` is \
+ used.
:param same_gene: Say if we consider as co-localised exon within the \
same gene
:param html_fig: True to create an html figure, false else
:param feature: The feature we want to analyse (default 'exon')
+ :param project: The project name, used only if global_weight = 0
:param logging_level: Level of information to display
"""
ConfigGraph.community_folder.mkdir(exist_ok=True, parents=True)
logging_def(ConfigGraph.community_folder, __file__, logging_level)
- global_weights = list(np.unique(global_weights))
- weights = list(np.unique(weights))
- projects, dic_project = get_projects_name(global_weights)
- condition = list(product(projects, weights))
- processes = []
- pool = mp.Pool(processes=min(ps, len(condition)))
- for project, weight in condition:
- global_weight = dic_project[project]
- logging.info(f'Finding community for project : {project}, '
- f'global_weight : {global_weight}, weight: {weight}')
- args = [weight, global_weight, project, same_gene, html_fig, feature]
- processes.append(pool.apply_async(community_finder, args))
- for proc in processes:
- proc.get(timeout=None)
- pool.close()
- pool.join()
+ project = get_projects(global_weight, project)
+ logging.info(f'Finding community for project : {project}, '
+ f'global_weight : {global_weight}, weight: {weight}')
+ community_finder(weight, global_weight, project, same_gene, html_fig,
+ feature)
diff --git a/src/find_interaction_cluster/create_ppi_files.py b/src/find_interaction_cluster/create_ppi_files.py
index 0c7a211d..1d401736 100644
--- a/src/find_interaction_cluster/create_ppi_files.py
+++ b/src/find_interaction_cluster/create_ppi_files.py
@@ -12,7 +12,7 @@ from typing import List
from ..logging_conf import logging_def
import numpy as np
import logging
-from .community_finder import get_projects_name
+from .community_finder import get_projects
from itertools import product
import multiprocessing as mp
from pathlib import Path
@@ -195,7 +195,8 @@ def filter_most_overllaping_ppi(df: pd.DataFrame, size_threshold: int
:param df: The dataframe containing the number of \
genes that clusters in community at DNA and protein level
- :param the minimum size required to keep gene community at dna level.
+ :param size_threshold: the minimum size required to keep gene
+ community at dna level.
:return: The dataframe with only one line per DNA community
>>> test_df = pd.DataFrame({"id_gene": range(1, 12),
@@ -203,7 +204,7 @@ def filter_most_overllaping_ppi(df: pd.DataFrame, size_threshold: int
... "community_size": [9] * 5 + [14] * 5 + [19],
... "nb_com-ppi": [0, 1, 2, 2, 0, 0, 1, 0, 0, 0, 0],
... "size_com-ppi": [30, 50, 105, 102, 25, 30, 42, 47, 89, 12, 0]})
- >>> filter_most_overllaping_ppi(test_df)
+ >>> filter_most_overllaping_ppi(test_df, 10)
DNA_community community_size nb_com-ppi size_com-ppi
0 C1 10 3 105
1 C2 15 2 42
@@ -324,7 +325,7 @@ def ppi_stats_analysis(community_file: Path, fasterdb_ppi: Path,
df.to_csv(outstat, sep="\t", index=False)
final_df = summary_randomisation_test(df, dic_dna_gene, ppi_gene,
iteration, dic_values,
- use_seed = False)
+ use_seed=False)
outstat = ConfigGraph.get_community_file(project, weight, global_weight,
same_gene, "gene",
f"ppi_gene_table_{iteration}_"
@@ -333,8 +334,9 @@ def ppi_stats_analysis(community_file: Path, fasterdb_ppi: Path,
final_df.to_csv(outstat, sep="\t", index=False)
-def ppi_stat_launcher(ps: int, weights: List[int],
- global_weights: List[int],
+def ppi_stat_launcher(weight: int,
+ global_weight: int,
+ project: str,
same_gene: bool, threshold: int = 700,
iteration: int = 1000,
logging_level: str = "DISABLE"):
@@ -342,12 +344,11 @@ def ppi_stat_launcher(ps: int, weights: List[int],
Launch the statistical allowing to determine if interaction between \
genes at dna level as an influence on the interactions at protein level.
- :param ps: The number of processes we want to use.
- :param weights: The list of weights of interaction to consider
- :param global_weights: The list global weights to consider. if \
- the global weight is equal to 0 then then density figure are calculated \
- by project, else all projcet are merge together and the interaction \
- seen in `global_weight` project are taken into account
+ :param weight: The weight of interaction to consider
+ :param global_weight: The global weighs to consider. if \
+ the global weight is equal to 0 then the project `project` is \
+ used.
+ :param project: The project name, used only if global_weight = 0
:param same_gene: Say if we consider as co-localised exon within the \
same gene
:param threshold: The minimum threshold needed to consider the interaction
@@ -358,24 +359,15 @@ def ppi_stat_launcher(ps: int, weights: List[int],
logging_def(ConfigGraph.community_folder, __file__, logging_level)
logging.info("Checking if gene communities at DNA level have an "
"influence on communities at protein level")
- global_weights = list(np.unique(global_weights))
- weights = list(np.unique(weights))
- projects, dic_project = get_projects_name(global_weights)
- condition = list(product(projects, weights))
- pool = mp.Pool(processes=min(ps, len(condition)))
+ project = get_projects(global_weight, project)
logging.debug("Calculating stats...")
- processes = []
- for project, weight in condition:
- global_weight = dic_project[project]
- community_file = ConfigGraph.get_community_file(project, weight,
- global_weight,
- same_gene, "gene",
- f".txt")
- args = [community_file, ConfigPPI.fasterdb_ppi, project,
- weight, global_weight, same_gene, threshold, iteration]
- processes.append(pool.apply_async(ppi_stats_analysis, args))
- results = [p.get(timeout=None) for p in processes]
+ community_file = ConfigGraph.get_community_file(project, weight,
+ global_weight,
+ same_gene, "gene",
+ f".txt")
+ ppi_stats_analysis(community_file, ConfigPPI.fasterdb_ppi, project,
+ weight, global_weight, same_gene, threshold, iteration)
if __name__ == "__main__":
- testmod()
\ No newline at end of file
+ testmod()
diff --git a/src/find_interaction_cluster/nt_and_community.py b/src/find_interaction_cluster/nt_and_community.py
index 0b7a0140..8e7b6394 100644
--- a/src/find_interaction_cluster/nt_and_community.py
+++ b/src/find_interaction_cluster/nt_and_community.py
@@ -18,8 +18,7 @@ from functools import reduce
from pathlib import Path
from rpy2.robjects import r, pandas2ri
from statsmodels.stats.multitest import multipletests
-import numpy as np
-from .community_finder import get_projects_name
+from .community_finder import get_projects
from ..logging_conf import logging_def
from itertools import product
import multiprocessing as mp
@@ -64,8 +63,8 @@ def get_nt_frequency(cnx: sqlite3.Connection, list_ft: List[str],
{query_region}
"""
df = pd.read_sql_query(query, cnx)
- df = df.pivot_table(index=f"id_{feature}", columns="ft", values="frequency")\
- .reset_index()
+ df = df.pivot_table(index=f"id_{feature}", columns="ft",
+ values="frequency").reset_index()
df[f"id_{feature}"] = df[f"id_{feature}"].astype(str)
return df
@@ -203,6 +202,8 @@ def create_ctrl_community(df: pd.DataFrame, outfile: Path,
community.
:param outfile: The output table containing frequencies
:param feature: The kind of feature to analyse
+ :param region: only use if feature is 'gene'. Used to focus on \
+ a given region in genes (can be gene, exon, intron).
:return: A dataframe containing the frequency of every nucleotides \
of every exon in a large community
"""
@@ -304,20 +305,22 @@ def create_dataframe(project: str, weight: int, global_weight: int,
return df
-def multiple_nt_lmm_launcher(ps: int, weights: List[int],
- global_weights: List[int],
+def multiple_nt_lmm_launcher(ps: int,
+ weight: int,
+ global_weight: int,
+ project: str,
same_gene: bool,
feature: str = 'exon', region: str = '',
logging_level: str = "DISABLE"):
"""
Launch the statistical analysis for every
- :param ps: The number of processes we want to use.
- :param weights: The list of weights of interaction to consider
- :param global_weights: The list global weights to consider. if \
- the global weight is equal to 0 then then density figure are calculated \
- by project, else all projcet are merge together and the interaction \
- seen in `global_weight` project are taken into account
+ :param ps: The number of processes to use
+ :param weight: The weight of interaction to consider
+ :param global_weight: The global weighs to consider. if \
+ the global weight is equal to 0 then the project `project` is \
+ used.
+ :param project: The project name, used only if global_weight = 0
:param same_gene: Say if we consider as co-localised exon within the \
same gene
:param feature: The kind of analysed feature
@@ -328,17 +331,14 @@ def multiple_nt_lmm_launcher(ps: int, weights: List[int],
logging_def(ConfigGraph.community_folder, __file__, logging_level)
logging.info("Checking if communities as an effect on nucleotide "
"frequency")
- global_weights = list(np.unique(global_weights))
- weights = list(np.unique(weights))
- projects, dic_project = get_projects_name(global_weights)
+ project = get_projects(global_weight, project)
nt_list = ["A", "C", "G", "T", "S", "W"]
- condition = list(product(projects, weights, nt_list))
+ condition = list(product([project], [weight], nt_list))
processes = {}
pool = mp.Pool(processes=min(ps, len(condition)))
logging.debug("Calculating stats...")
dic_df = {}
for project, weight, nt in condition:
- global_weight = dic_project[project]
ckey = get_key(project, weight)
if ckey in dic_df:
df = dic_df[ckey]
@@ -352,11 +352,11 @@ def multiple_nt_lmm_launcher(ps: int, weights: List[int],
dic_df[ckey] = df
args = [df, project, weight, global_weight, same_gene, nt, feature,
region]
- if ckey not in processes.keys():
- processes[ckey] = [pool.apply_async(get_stat_nt_communities, args)]
- else:
+ if ckey in processes:
processes[ckey].append(
pool.apply_async(get_stat_nt_communities, args))
+ else:
+ processes[ckey] = [pool.apply_async(get_stat_nt_communities, args)]
for p, value in processes.items():
project, weight = p.split("_")
results = [p.get(timeout=None) for p in value]
@@ -365,7 +365,7 @@ def multiple_nt_lmm_launcher(ps: int, weights: List[int],
fdf = pd.DataFrame(results)
fdf["padj"] = multipletests(fdf['pval'].values, method='fdr_bh')[1]
outfile = ConfigGraph.get_community_file(project, weight,
- dic_project[project],
+ global_weight,
same_gene, feature,
f"lmm-nt_stat.txt",
"sf_community_enrichment")
diff --git a/src/find_interaction_cluster/sf_and_communities.py b/src/find_interaction_cluster/sf_and_communities.py
index ac965851..d81dc593 100644
--- a/src/find_interaction_cluster/sf_and_communities.py
+++ b/src/find_interaction_cluster/sf_and_communities.py
@@ -15,7 +15,7 @@ from .community_finder import get_communities
import pandas as pd
import numpy as np
from itertools import product
-from .community_finder import get_projects_name
+from .community_finder import get_projects
import multiprocessing as mp
import logging
from ..logging_conf import logging_def
@@ -362,19 +362,21 @@ def get_key(project: str, weight: int) -> str:
return f"{project}_{weight}"
-def multiple_stat_launcher(ps: int, weights: List[int],
- global_weights: List[int],
+def multiple_stat_launcher(ps: int,
+ weight: int,
+ global_weight: int,
+ project: str,
same_gene: bool, feature: str = 'exon',
logging_level: str = "DISABLE"):
"""
Launch the statistical analysis for every
- :param ps: The number of processes we want to use.
- :param weights: The list of weights of interaction to consider
- :param global_weights: The list global weights to consider. if \
- the global weight is equal to 0 then then density figure are calculated \
- by project, else all projcet are merge together and the interaction \
- seen in `global_weight` project are taken into account
+ :param ps: The number of processes to use
+ :param weight: The weight of interaction to consider
+ :param global_weight: The global weighs to consider. if \
+ the global weight is equal to 0 then the project `project` is \
+ used.
+ :param project: The project name, used only if global_weight = 0
:param same_gene: Say if we consider as co-localised exon within the \
same gene
:param feature: The feature we want to analyse
@@ -385,22 +387,19 @@ def multiple_stat_launcher(ps: int, weights: List[int],
logging.info(f"Checking if communities contains often {feature}s "
f"regulated by a splicing factor")
sf_list = get_sfname()
- global_weights = list(np.unique(global_weights))
- weights = list(np.unique(weights))
- projects, dic_project = get_projects_name(global_weights)
- condition = list(product(projects, weights, sf_list, ['down', 'up']))
+ project = get_projects(global_weight, project)
+ condition = list(product([project], [weight], sf_list, ['down', 'up']))
processes = {}
pool = mp.Pool(processes=min(ps, len(condition)))
logging.debug("Calculating stats...")
for project, weight, sf_name, reg in condition:
ckey = get_key(project, weight)
- global_weight = dic_project[project]
args = [sf_name, reg, project, weight, global_weight, same_gene,
feature]
- if ckey not in processes.keys():
- processes[ckey] = [pool.apply_async(get_stat4communities, args)]
- else:
+ if ckey in processes:
processes[ckey].append(pool.apply_async(get_stat4communities, args))
+ else:
+ processes[ckey] = [pool.apply_async(get_stat4communities, args)]
for p, value in processes.items():
project, weight = p.split("_")
list_tuples = [proc.get(timeout=None) for proc in value]
@@ -410,7 +409,7 @@ def multiple_stat_launcher(ps: int, weights: List[int],
list_series = [t[1] for t in list_tuples]
df = pd.concat(list_df, axis=0, ignore_index=True)
outfile = ConfigGraph.get_community_file(project, weight,
- dic_project[project],
+ global_weight,
same_gene, feature,
"_stat.txt",
"sf_community_enrichment")
@@ -419,7 +418,7 @@ def multiple_stat_launcher(ps: int, weights: List[int],
glm_df["padj"] = multipletests(glm_df['pval'].values,
method='fdr_bh')[1]
outfile = ConfigGraph.get_community_file(project, weight,
- dic_project[project],
+ global_weight,
same_gene, feature,
"_glmm_stat.txt",
"sf_community_enrichment")
--
GitLab