Tutorial 4¶
1. Import Data¶
- Import the
pandaspython module and name itpd - Load the Excel sheet "Data" from Excel file "data.xlsx" into a dataframe called
data. - Display the number of rows and columns in
data. - Display the first 10 rows at the top (head) of the
datatable. - Load the Excel sheet "Peak" from Excel file "data.xlsx" into a dataframe called
peak. - Display the number of rows and columns in
peak. - Display the first 10 rows at the top (head) of the
peaktable.
In [6]:
import pandas as pd
data = pd.read_excel('data.xlsx', sheet_name='Data')
print("Data Table: {} rows & {} columns".format(*data.shape))
display(data.head(10)) # View data table (top 10 rows)
peak = pd.read_excel('data.xlsx', sheet_name='Peak')
print("Peak Table: {} rows & {} columns".format(*peak.shape))
display(peak.head(10))
2. Histogram of RSD¶
- Import the
matplotlib.pyplotvisualisation package and name itplt - Invoke the special "Jupyter Magic" command
%matplotlib inlineto help simplify plotting - Create a histogram of the column
peak.RSD - Add an x-axis label "RSD" to the figure
- Generate the figure using the function
plt.show()
In [2]:
import matplotlib.pyplot as plt
%matplotlib inline
plt.hist(peak.RSD, 50, density=True, facecolor='g', alpha=0.5)
plt.xlabel('RSD', fontsize=15)
plt.show()
3. Jointplot of RSD vs. D-Ratio¶
- Import the
seabornvisualisation package and name itsns - Create a plot of
peakcolumnspeak.RSDvs.peak.D-Ratiowith bivariate and distribution graphs. (automatically displayed)
In [3]:
import seaborn as sns
sns.jointplot(x=peak.RSD, y=peak.D_Ratio, kind='kde', color="skyblue")
Out[3]:
4. PCA score plot of QC vs. Sample¶
- Import the packages required to calculate the principal component analysis (PCA)
numpy(import asnp)scikit-learnpackagessklearn.decomposition.PCA,sklearn.preprocessing.StandardScaler
- Import the matplotlib package required to map colours to datapoints, to plot the PCA scores
matplotlib.colors.ListedColormap
- Scale the data and calculate the PCA.
- Plot PCA scores (pc1 vs pc2) labeling/coloring samples by 'QC' or 'Sample'
In [4]:
# Import
import numpy as np
from sklearn.decomposition import PCA
from sklearn.preprocessing import StandardScaler
from matplotlib.colors import ListedColormap
# Extract X matrix
names = peak['Name']
x = data[names].values
x = np.log(x)
x = StandardScaler().fit_transform(x)
# Create and fit PCA
pca = PCA(n_components=2)
scores = pca.fit_transform(x)
label = data['SampleType']
# Split scores into sample and QC
Sample_scores = scores[label == 'Sample',:]
QC_scores = scores[label == 'QC',:]
# Plot Sample score and QC score
fig = plt.figure(figsize=(8,8))
h1 = plt.scatter(Sample_scores[:,0],Sample_scores[:,1],edgecolors='Black', facecolors='Green',s=100,alpha=0.5)
h2 = plt.scatter(QC_scores[:,0],QC_scores[:,1], edgecolors='Black', facecolors='Red',s=100,alpha=0.5)
# Add legend, labels, and title
plt.legend((h1,h2),('Sample','QC'),fontsize=15)
plt.xlabel('PC1', fontsize=15)
plt.ylabel('PC2', fontsize=15)
plt.title('Quality Control PCA plot',fontsize=20)
# Show plot
plt.show()
5. Bubble plot of Molecular Weights vs. Retention Time (sized by RSD)¶
- Create a scatterplot of two
peakcolumns:peak.Mol_Weightvs.peak.RT_minutes(each dot is a metabolite peak).- make the size of the dots proportional to the RSD of the datapoint with the argument
s=peak.RSD**2/2 - make the dots (bubbles) transparent with the argument
alpha=0.2
- make the size of the dots proportional to the RSD of the datapoint with the argument
- Add x-axis label "Molecular Weight", and y-axis label "RT minutes"
- Add a plot title
- Render the plot
In [5]:
# Scatterplot of Mol_Weight vs. RT_minute with size RSD^2/2, and colour red
fig = plt.figure(figsize=(20,16))
plt.scatter(peak.Mol_Weight, peak.RT_minutes, s=peak.RSD**2/2, alpha=0.2, edgecolors='black', c='red')
plt.xlabel('Molecular Weight', fontsize=15)
plt.ylabel('RT minutes', fontsize=15)
plt.title('Metabolites Detected (sized by RSD)',fontsize=20)
plt.show()
In [ ]:
In [ ]: