Skip to content

redo mAP and figure4 #166

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Mar 12, 2025
Merged

redo mAP and figure4 #166

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e98d439
redo mAP and figure4
MikeLippincott Mar 11, 2025
a09feac
addressing comments and finalizing plots
MikeLippincott Mar 12, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
415 changes: 333 additions & 82 deletions ...lk_Morphology_Elastic_Network/2.test_model/notebooks/1.test_regression_multi_output.ipynb
View file
Open in desktop

Large diffs are not rendered by default.

Empty file removed ...ology_Elastic_Network/2.test_model/results/regression/PBMC_aggregated_with_nomic/.gitkeep
View file
Open in desktop
Empty file.
173 changes: 90 additions & 83 deletions 6.bulk_Morphology_Elastic_Network/2.test_model/scripts/1.test_regression_multi_output.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
#!/usr/bin/env python
# coding: utf-8

# In[ ]:
# In[1]:


import argparse
Expand All @@ -16,7 +16,7 @@
from sklearn.metrics import explained_variance_score, mean_squared_error, r2_score
from sklearn.utils import parallel_backend

# In[ ]:
# In[2]:


argparser = argparse.ArgumentParser()
Expand All @@ -31,16 +31,20 @@
cytokine = args.cytokine
print(cell_type, shuffle, cytokine)

# cell_type = "PBMC"
# shuffle = False
# cytokine = "XCL1 (Lymphotactin) [NSU]"

# In[ ]:

# In[3]:


# Parameters
aggregation = True
nomic = True


# In[ ]:
# In[4]:


# set shuffle value
Expand All @@ -50,13 +54,13 @@
shuffle = "final"


# In[ ]:
# In[5]:


MODEL_TYPE = "regression"


# In[ ]:
# In[6]:


# load training data from indexes and features dataframe
Expand All @@ -73,7 +77,7 @@
data_split_indexes = pd.read_csv(data_split_path, sep="\t")


# In[ ]:
# In[7]:


# rename column that contain the treatment dose to be a metadata column
Expand All @@ -85,30 +89,30 @@
)


# In[ ]:
# In[8]:


# remove duplicate columns
data_df = data_df.loc[:, ~data_df.columns.duplicated()]


# In[ ]:
# In[9]:


# select tht indexes for the training and test set
train_indexes = data_split_indexes.loc[data_split_indexes["label"] == "train"]
test_indexes = data_split_indexes.loc[data_split_indexes["label"] == "test"]


# In[ ]:
# In[10]:


# subset data_df by indexes in data_split_indexes
training_data = data_df.loc[train_indexes["labeled_data_index"]]
testing_data = data_df.loc[test_indexes["labeled_data_index"]]


# In[ ]:
# In[11]:


# define metadata columns
Expand All @@ -135,29 +139,29 @@
test_data_x = test_data_x.drop(test_data_y_cols, axis=1)


# In[ ]:
# In[12]:


print(train_data_x.shape, train_data_y.shape, test_data_x.shape, test_data_y.shape)


# In[ ]:
# In[13]:


# set model path from parameters
if (aggregation == True) and (nomic == True):
model_path = pathlib.Path(f"models/regression/{cell_type}/aggregated_with_nomic/")
model_path = pathlib.Path(f"models/regression/{cell_type}_aggregated_with_nomic/")
elif (aggregation == True) and (nomic == False):
model_path = pathlib.Path(f"models/regression/{cell_type}/aggregated/")
model_path = pathlib.Path(f"models/regression/{cell_type}_aggregated/")
elif (aggregation == False) and (nomic == True):
model_path = pathlib.Path(f"models/regression/{cell_type}/sc_with_nomic/")
model_path = pathlib.Path(f"models/regression/{cell_type}_sc_with_nomic/")
elif (aggregation == False) and (nomic == False):
model_path = pathlib.Path(f"models/regression/{cell_type}/sc/")
model_path = pathlib.Path(f"models/regression/{cell_type}_sc/")
else:
print("Error")


# In[ ]:
# In[14]:


data_dict = {
Expand All @@ -176,14 +180,14 @@
}


# In[ ]:
# In[15]:


# list of metrics to use
output_metric_scores = {}


# In[ ]:
# In[16]:


# blank df for concatenated results
Expand All @@ -207,71 +211,74 @@
)


# In[ ]:
# In[17]:


list_of_dfs = []
for data_split in data_dict.keys():
data_x = data_dict[data_split]["data_x"]
data_y = data_dict[data_split]["data_y"]
col_names = data_dict[data_split]["col_names"]
metadata = data_dict[data_split]["metadata"]

if shuffle == "shuffled_baseline":
model = joblib.load(
f"../../1.train_models/{model_path}/{cytokine}_shuffled_baseline__all_nomic.joblib"
)
elif shuffle == "final":
model = joblib.load(
f"../../1.train_models/{model_path}/{cytokine}_final__all_nomic.joblib"
)
else:
print("Error")

# get the cytokine column of choice
y_selected = data_y[cytokine]

if shuffle == "shuffled_baseline":
for column in data_x:
np.random.shuffle(data_x[column].values)

# get predictions
predictions = model.predict(data_x)

explained_variance = explained_variance_score(y_selected, predictions)
output_metric_scores["explained_variance"] = explained_variance
neg_mean_absolute_error = -mean_squared_error(y_selected, predictions)
output_metric_scores["neg_mean_absolute_error"] = neg_mean_absolute_error
neg_mean_squared_error = -mean_squared_error(y_selected, predictions)
output_metric_scores["neg_mean_squared_error"] = neg_mean_squared_error
r2 = r2_score(y_selected, predictions)
output_metric_scores["treatment"] = metadata[
"oneb_Metadata_Treatment_Dose_Inhibitor_Dose"
]
output_metric_scores["well"] = metadata["Metadata_Well"].values
df = pd.DataFrame.from_dict(output_metric_scores)
df["r2"] = r2
df["cytokine"] = cytokine
df["data_split"] = data_split
df["shuffle"] = shuffle
df["predicted_value"] = predictions
df["actual_value"] = y_selected
df["log10_neg_mean_absolute_error"] = -np.log10(-df["neg_mean_absolute_error"])
df["log10_neg_mean_squared_error"] = -np.log10(-df["neg_mean_squared_error"])
df["log10_explained_variance"] = -np.log10(df["explained_variance"])

# replace "[NSU]" with """
df["cytokine"] = df["cytokine"].replace("[ \[\]NSU]", "", regex=True)
df["cytokine"] = df["cytokine"].replace(" ", "_", regex=True)
list_of_dfs.append(df)
# concat the dataframes
results_df = pd.concat(list_of_dfs, axis=0)

data_x = test_data_x
data_y = test_data_y
metadata = metadata_test
data_split = "test"
if shuffle == "shuffled_baseline":
model = joblib.load(
f"../../1.train_models/{model_path}/{cytokine}_shuffled_baseline__all_nomic.joblib"
)
elif shuffle == "final":
model = joblib.load(
f"../../1.train_models/{model_path}/{cytokine}_final__all_nomic.joblib"
)
else:
print("Error")

# get the cytokine column of choice
y_selected = data_y[cytokine]

if shuffle == "shuffled_baseline":
for column in data_x:
np.random.shuffle(data_x[column].values)

# get predictions
predictions = model.predict(data_x)

explained_variance = explained_variance_score(y_selected, predictions)
output_metric_scores["explained_variance"] = explained_variance
neg_mean_absolute_error = -mean_squared_error(y_selected, predictions)
output_metric_scores["neg_mean_absolute_error"] = neg_mean_absolute_error
neg_mean_squared_error = -mean_squared_error(y_selected, predictions)
output_metric_scores["neg_mean_squared_error"] = neg_mean_squared_error
r2 = r2_score(y_selected, predictions)
output_metric_scores["treatment"] = metadata[
"oneb_Metadata_Treatment_Dose_Inhibitor_Dose"
]
output_metric_scores["well"] = metadata["Metadata_Well"].values
df = pd.DataFrame.from_dict(output_metric_scores)
df["r2"] = r2
df["cytokine"] = cytokine
df["data_split"] = data_split
df["shuffle"] = shuffle
df["predicted_value"] = predictions
df["actual_value"] = y_selected
df["log10_neg_mean_absolute_error"] = -np.log10(-df["neg_mean_absolute_error"])
df["log10_neg_mean_squared_error"] = -np.log10(-df["neg_mean_squared_error"])
df["log10_explained_variance"] = -np.log10(df["explained_variance"])

# replace "[NSU]" with """
df["cytokine"] = df["cytokine"].replace("[ \[\]NSU]", "", regex=True)
df["cytokine"] = df["cytokine"].replace(" ", "_", regex=True)

# concat the dataframes
results_df = pd.concat([results_df, df], axis=0)


# In[ ]:
# In[18]:


results_df.head()


# In[ ]:
# In[19]:


var_df = results_df.drop(
Expand Down Expand Up @@ -299,26 +306,26 @@
var_df.head()


# In[ ]:
# In[20]:


# set model path from parameters
if (aggregation == True) and (nomic == True):
results_path = pathlib.Path(
f"../results/regression/{cell_type}/aggregated_with_nomic/"
f"../results/regression/{cell_type}_aggregated_with_nomic/"
)
elif (aggregation == True) and (nomic == False):
results_path = pathlib.Path(f"../results/regression/{cell_type}/aggregated/")
results_path = pathlib.Path(f"../results/regression/{cell_type}_aggregated/")
elif (aggregation == False) and (nomic == True):
results_path = pathlib.Path(f"../results/regression/{cell_type}/sc_with_nomic/")
results_path = pathlib.Path(f"../results/regression/{cell_type}_sc_with_nomic/")
elif (aggregation == False) and (nomic == False):
results_path = pathlib.Path(f"../results/regression/{cell_type}/sc/")
results_path = pathlib.Path(f"../results/regression/{cell_type}_sc/")
else:
print("Error")
pathlib.Path(results_path).mkdir(parents=True, exist_ok=True)


# In[ ]:
# In[21]:


# check if the model training metrics file exists
Expand Down
36 changes: 26 additions & 10 deletions 6.bulk_Morphology_Elastic_Network/2.test_model/scripts/2.combine_regression_tests.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4,35 +4,36 @@
# In[1]:


import argparse
import pathlib

import pandas as pd

# In[2]:


argparser = argparse.ArgumentParser()
argparser.add_argument("--cell_type", type=str, default="cell_type")
# argparser = argparse.ArgumentParser()
# argparser.add_argument("--cell_type", type=str, default="cell_type")

args = argparser.parse_args()
# args = argparser.parse_args()

cell_type = args.cell_type
# cell_type = args.cell_type

cell_type = "PBMC"

# In[ ]:

# In[3]:


results_dir_path = pathlib.Path(
f"../results/regression/{cell_type}/aggregated_with_nomic/"
f"../results/regression/{cell_type}_aggregated_with_nomic/"
).resolve(strict=True)

model_stats_final_output_path = pathlib.Path(
f"../results/regression/{cell_type}/aggregated_with_nomic/model_stats.csv"
f"../results/regression/{cell_type}_aggregated_with_nomic/model_stats.csv"
)

variance_r2_stats_final_output_path = pathlib.Path(
f"../results/regression/{cell_type}/aggregated_with_nomic/variance_r2_stats.csv"
f"../results/regression/{cell_type}_aggregated_with_nomic/variance_r2_stats.csv"
)

# get a list of all the files that contain "model_stats" in the name
Expand All @@ -54,7 +55,22 @@
variance_r2_stats_df.to_csv(variance_r2_stats_final_output_path, index=False)


# In[ ]:
# In[4]:


print("Completed!")


# In[5]:


model_stats_df.head()


# In[6]:


model_stats_df["data_split"].unique()


# In[ ]:
Loading