Per a la creació d'aquesta recerca, hem descarregat dades de la pàgina RCSB Protein Data Bank per comprovar la seva efectivitat contra el càncer
la meva proteïna és la 3QFP la qual és una chaperona (chaperone).
Per entendre-ho millor les chaperones moleculars són proteïnes que ajuden al plegament conformacional o al desplegament de grans proteïnes o complexos de proteïnes macromoleculars. Hi ha diverses classes de chaperones moleculars, totes les quals funcionen per ajudar les proteïnes grans a plegar-se adequadament durant o després de la síntesi i després de la desnaturalització parcial. Els chaperons també estan implicats en la translocació de proteïnes per a la proteòlisi.
gràcies a les dades tretes d'aquests gràfics podem concloure que la Taxinine J és la millor molècula a l'hora de reaccionar amb més efectivitat a la proteïna 3QFP
Abans de posar el codi expliquem que està fet amb Google Colab, una biblioteca on es pot triar un entorn d'execució Python o R i GPU o CPU, entrant a Edit-Notebook settings. La GPU és la unitat de processament gràfic i la CPU és la unitat de processament del computador. La GPU és més útil en DeepLearning, perquè permet processament simultani de codi (en paral·lel). La CPU pot tenir diversos nuclis, però és més lenta en DeepLearning o visió per ordinador.
# 2. Upload deltaG_values.csv of every docking to generate a boxplot to obtain a summary table transposed data in csv and boxplot
# Has de pujar els diferents arxius i penjar cancel quan acabis, posar els noms en català (en aquest exemple: Taxinina A, etc)
# Després has d'escriure el codi PDB de la teva proteïna (en aquest exemple 4HFZ)
from google.colab import files
import pandas as pd
import matplotlib.pyplot as plt
import io
import numpy as np
# Initialize an empty list to store DataFrame objects
dfs = []
# Upload CSV files one by one
print("Upload CSV files one by one. Press Cancel to stop uploading.")
while True:
uploaded_files = files.upload()
if len(uploaded_files) == 0:
break
for filename, contents in uploaded_files.items():
# Read CSV file as DataFrame and append it to the list
df = pd.read_csv(io.StringIO(contents.decode('utf-8')), header=None)
# Add a column to identify the compound
df['Compound'] = f'Compound {chr(ord("A") + len(dfs))}'
dfs.append(df)
# Concatenate DataFrames vertically
combined_df = pd.concat(dfs, ignore_index=True)
# Transpose the DataFrame so that rows become columns
transposed_df = combined_df.set_index('Compound').T
# Save the transposed DataFrame to a new CSV file
transposed_csv_path = 'transposed_data.csv'
transposed_df.to_csv(transposed_csv_path)
# Prompt the user to enter real chemical names for each compound
real_names_mapping = {}
for i, df_name in enumerate(transposed_df.columns):
real_name = input(f"Enter the real chemical name for {df_name}: ")
real_names_mapping[df_name] = real_name
# Prompt the user to enter the last word of the graph title
graph_title_suffix = input("Enter the last word of the graph title: ").strip()
# Create a customized boxplot for compounds
plt.figure(figsize=(8, 6))
# Set colors
box_color = 'blue'
median_color = 'orange'
whisker_color = 'green'
cap_color = 'purple'
# Create a boxplot
boxprops = dict(color=box_color)
medianprops = dict(color=median_color)
whiskerprops = dict(color=whisker_color)
capprops = dict(color=cap_color)
boxplot = transposed_df.boxplot(vert=False, return_type='dict', boxprops=boxprops, medianprops=medianprops, whiskerprops=whiskerprops, capprops=capprops)
# Overlay individual data points
for df_name in transposed_df.columns:
y = np.random.normal(list(transposed_df.columns).index(df_name) + 1, 0.1, size=len(transposed_df[df_name]))
plt.scatter(transposed_df[df_name], y, alpha=0.5, s=10)
# Set ticks and labels
plt.yticks(np.arange(1, len(transposed_df.columns) + 1), [real_names_mapping[col] for col in transposed_df.columns])
plt.xlabel("Energia d'unió (kcal/mol)")
plt.ylabel("Lligands")
plt.title(f"Acoblament molecular amb proteïna PDB {graph_title_suffix}")
plt.grid(True)
plt.axvline(x=0, color='red', linestyle='--') # Add line at 0 for reference
plt.tight_layout()
# Save the plot as an image file
plot_image_path = 'boxplot.png'
plt.savefig(plot_image_path)
# Download the transposed CSV file and the plot image
files.download(transposed_csv_path)
files.download(plot_image_path)
# Print paths to the saved files
print("Transposed data saved to:", transposed_csv_path)
print("Plot image saved to:", plot_image_path)
# 3. Upload several deltaG_values.csv from different molecular dockings to obtain a bar type figure
from google.colab import files
import pandas as pd
import matplotlib.pyplot as plt
import io
# Initialize an empty list to store DataFrame objects
dfs = []
# Upload CSV files one by one
print("Upload CSV files one by one. Press Cancel to stop uploading.")
while True:
uploaded_files = files.upload()
if len(uploaded_files) == 0:
break
for filename, contents in uploaded_files.items():
# Read CSV file as DataFrame and append it to the list
df = pd.read_csv(io.StringIO(contents.decode('utf-8')), header=None)
# Add a column to identify the compound
df['Compound'] = f'Compound {chr(ord("A") + len(dfs))}'
dfs.append(df)
# Concatenate DataFrames vertically
combined_df = pd.concat(dfs, ignore_index=True)
# Transpose the DataFrame so that rows become columns
transposed_df = combined_df.set_index('Compound').T
# Save the transposed DataFrame to a new CSV file
transposed_csv_path = 'transposed_data.csv'
transposed_df.to_csv(transposed_csv_path)
# Prompt the user to enter customized names for each compound
custom_names_mapping = {}
for i, df_name in enumerate(transposed_df.columns):
custom_name = input(f"Enter customized name for {df_name}: ")
custom_names_mapping[df_name] = custom_name
# Calculate statistics for each compound
statistics_df = pd.DataFrame({
'Mean': transposed_df.mean(),
'Standard Deviation': transposed_df.std(),
'Minimum': transposed_df.min(),
'Maximum': transposed_df.max(),
'Number of Data': transposed_df.count()
})
# Save statistics to a CSV file
statistics_csv_path = 'compound_statistics.csv'
statistics_df.to_csv(statistics_csv_path)
# Create a bar graph with mean and standard deviation
plt.figure(figsize=(10, 6))
mean_values = statistics_df['Mean']
std_values = statistics_df['Standard Deviation']
plt.bar([custom_names_mapping[col] for col in mean_values.index], mean_values, yerr=std_values, capsize=5, color='skyblue', edgecolor='black')
plt.xlabel('Compound')
plt.ylabel('Binding Energy (kcal/mol)')
plt.title('Mean Binding Energy with Standard Deviation')
plt.xticks(rotation=45)
plt.grid(axis='y')
plt.tight_layout()
# Save the plot as an image file
plot_image_path = 'mean_with_std_bar_plot.png'
plt.savefig(plot_image_path)
# Download the transposed DataFrame, statistics CSV file, and plot image
files.download(transposed_csv_path)
files.download(statistics_csv_path)
files.download(plot_image_path)
# Display the statistics DataFrame
statistics_df