diff --git a/src/nt_composition/make_nt_correlation.py b/src/nt_composition/make_nt_correlation.py
index d0e13294736fa23602662a16813b9d4025c92771..286daa528b0ea3f34777fac8c755909b58e8a5eb 100644
--- a/src/nt_composition/make_nt_correlation.py
+++ b/src/nt_composition/make_nt_correlation.py
@@ -169,25 +169,23 @@ def get_dic_co_regulated_exon(arr_interaction: np.array):
return d
-def create_density_table(arr_interaction: np.array, dic_freq: Dict[str, float],
- ) -> pd.DataFrame:
+def density_mean(arr_interaction: np.array, dic_freq: Dict[str, float]):
"""
Create the density table, a table showing the frequency of \
- a nucleotide in every exon in a chia-pet projet and the mean \
- frequency of every other co-localised exons.
+ a nucleotide in every exon in a chia-pet project and the mean \
+ frequency of every other co-localised exons
:param arr_interaction: array of interaction between exons.
:param dic_freq: The frequency dataframe.
:return: The density table
"""
- logging.debug(f'Calculating density table ({os.getpid()})')
exons_dic = get_dic_co_regulated_exon(arr_interaction)
dic = {'exon': [], 'freq_exon': [], 'freq_coloc_exon': [], 'oexon': []}
for exon in exons_dic.keys():
freq_ex = dic_freq[exon]
oexon = np.unique(exons_dic[exon])
freq_oexon = np.nanmean(np.asarray([dic_freq[ex] for ex in oexon],
- dtype=float))
+ dtype=float))
if freq_ex is not None and not np.isnan(freq_oexon):
dic['exon'].append(exon)
dic['freq_exon'].append(freq_ex)
@@ -198,6 +196,51 @@ def create_density_table(arr_interaction: np.array, dic_freq: Dict[str, float],
return df
+def simple_density(arr_interaction: np.array, dic_freq: Dict[str, float]):
+ """
+ Create the density table, a table showing the frequency of \
+ a nucleotide in every exon in a chia-pet project and the \
+ frequency of of one co-localised exon.
+
+ :param arr_interaction: array of interaction between exons.
+ :param dic_freq: The frequency dataframe.
+ :return: The density table
+ """
+ dic = {'exon1': [], 'freq_exon': [], 'freq_coloc_exon': [], 'exon2': []}
+ for exon1, exon2 in arr_interaction:
+ freq_ex = dic_freq[exon1]
+ freq_ex2 = dic_freq[exon2]
+ if freq_ex is not None and freq_ex2 is not None:
+ dic['exon1'].append(exon1)
+ dic['freq_exon'].append(freq_ex)
+ dic['freq_coloc_exon'].append(freq_ex2)
+ dic['exon2'].append(exon2)
+ df = pd.DataFrame(dic)
+ logging.debug(df.head())
+ return df
+
+
+def create_density_table(arr_interaction: np.array, dic_freq: Dict[str, float],
+ compute_mean: bool = False) -> pd.DataFrame:
+ """
+ Create the density table, a table showing the frequency of \
+ a nucleotide in every exon in a chia-pet project and the mean \
+ frequency of every other co-localised exons or the frequency of one \
+ co-localised exons.
+
+ :param arr_interaction: array of interaction between exons.
+ :param dic_freq: The frequency dataframe.
+ :param compute_mean: True to compute the mean frequency of co-localised \
+ exons, false to only compute the frequency of one co-localized exons.
+ :return: The density table
+ """
+ logging.debug(f'Calculating density table ({os.getpid()})')
+ if compute_mean:
+ return density_mean(arr_interaction, dic_freq)
+ else:
+ return simple_density(arr_interaction, dic_freq)
+
+
def create_density_fig(df: pd.DataFrame, project: str, ft_type: str, ft: str,
weight: int, global_weight: int) -> Tuple[float, float]:
"""
@@ -241,7 +284,7 @@ def create_density_fig(df: pd.DataFrame, project: str, ft_type: str, ft: str,
def create_density_figure(nt: str, ft_type: str,
project: str, weight: int, global_weight: int,
- same_gene: bool,
+ same_gene: bool, compute_mean: bool,
logging_level: str = "DISABLE"
) -> Tuple[float, float]:
"""
@@ -258,6 +301,8 @@ def create_density_figure(nt: str, ft_type: str,
seen in `global_weight` project are taken into account
:param same_gene: Say if we consider as co-localised exon within the \
same gene
+ :param compute_mean: True to compute the mean frequency of co-localised \
+ exons, false to only compute the frequency of one co-localized exons.
:param logging_level: The level of information to display
:return: The correlation and the p-value
"""
@@ -269,7 +314,7 @@ def create_density_figure(nt: str, ft_type: str,
arr_interaction = get_project_colocalisation(cnx, project, weight,
global_weight, same_gene)
dic_freq = get_frequency_dic(cnx, nt, ft_type)
- df = create_density_table(arr_interaction, dic_freq)
+ df = create_density_table(arr_interaction, dic_freq, compute_mean)
df.to_csv(outfile, sep="\t", index=False)
r, p = create_density_fig(df, project, ft_type, nt, weight,
global_weight)
@@ -320,7 +365,8 @@ def create_scatterplot(df_cor: pd.DataFrame, ft_type: str, ft: str,
def execute_density_figure_function(di: pd.DataFrame, project : str,
ft_type: str, ft: str, weight: int,
global_weight: int,
- same_gene: bool) -> Dict[str, Any]:
+ same_gene: bool,
+ compute_mean) -> Dict[str, Any]:
"""
Execute create_density_figure and organized the results in a dictionary.
@@ -335,11 +381,13 @@ def execute_density_figure_function(di: pd.DataFrame, project : str,
seen in `global_weight` project are taken into account
:param same_gene: Say if we consider as co-localised exon within the \
same gene
+ :param compute_mean: True to compute the mean frequency of co-localised \
+ exons, false to only compute the frequency of one co-localized exons.
:return:
"""
logging.info(f'Working on {project}, {ft_type}, {ft} - {os.getpid()}')
r, p = create_density_figure(ft, ft_type, project, weight,
- global_weight, same_gene)
+ global_weight, same_gene, compute_mean)
if global_weight == 0:
tmp = {"project": project, "ft_type": ft_type,
"ft": ft, "cor": r, "pval": p,
@@ -366,7 +414,7 @@ def combine_dic(list_dic: List[Dict]) -> Dict:
def create_all_frequency_figures(ps: int, weight: int = 1,
global_weight: int = 0, ft_type: str = "nt",
- same_gene = True,
+ same_gene = True, compute_mean: bool = True,
logging_level: str = "DISABLE"):
"""
Make density figure for every selected projects.
@@ -380,6 +428,8 @@ def create_all_frequency_figures(ps: int, weight: int = 1,
:param ft_type: The kind of feature to analyse
:param same_gene: Say if we consider as co-localised exon within the \
same gene
+ :param compute_mean: True to compute the mean frequency of co-localised \
+ exons, false to only compute the frequency of one co-localized exons.
:param ps: The number of processes to create
"""
logging_def(ConfigNt.interaction, __file__, logging_level)
@@ -397,7 +447,8 @@ def create_all_frequency_figures(ps: int, weight: int = 1,
pool = mp.Pool(processes=ps)
processes = []
for project, ft, ft_type in param:
- args = [di, project, ft_type, ft, weight, global_weight, same_gene]
+ args = [di, project, ft_type, ft, weight, global_weight, same_gene,
+ compute_mean]
processes.append(pool.apply_async(execute_density_figure_function,
args))
results = []