VCF file analysis method using Pandas

1. Data type conversion and filtering

Transform the data types of columns required for data analysis.

For example, POS can be converted to integer and QUAL to float.

# Data type conversion
df['POS'] = df['POS'].astype(int)
df['QUAL'] = df['QUAL'].astype(float)
df['AF'] = df['AF'].apply(lambda x: float(x.split(',')[0]))

# Filtering only high-quality variations
high_quality_variants = df[df['QUAL'] >= 30]
print(high_quality_variants)

2. High-quality variation analysis

It analyzes the frequency, location, genes, and so on of high quality variation.

# Calculation of frequency for each variant location
position_counts = high_quality_variants['POS'].value_counts()
print(position_counts)

# Calculation of ALT frequency
alt_freq = high_quality_variants['ALT'].value_counts()
print(alt_freq)

3. Visualization

Variant data are visualized using Matplotlib or Seaborn.

import matplotlib.pyplot as plt
import seaborn as sns

# Variant quality distribution visualization
plt.figure(figsize=(10, 6))
sns.histplot(high_quality_variants['QUAL'], bins=30, kde=True)
plt.title('Distribution of Variant Quality Scores')
plt.xlabel('Quality Score')
plt.ylabel('Frequency')
plt.show()

# Variant position distribution visualization
plt.figure(figsize=(10, 6))
sns.scatterplot(x='POS', y='QUAL', data=high_quality_variants)
plt.title('Variant Positions and Quality')
plt.xlabel('Position')
plt.ylabel('Quality Score')
plt.show()

4. One more step: Analysis of mutations within specific genes

Analyze and visualize variation for a particular gene.

# For example, filtering variation in a particular gene region.
gene_start = 10000
gene_end = 50000
gene_variants = high_quality_variants[(high_quality_variants['POS'] >= gene_start) & (high_quality_variants['POS'] <= gene_end)]

# Visualization of mutations within the gene area
plt.figure(figsize=(10, 6))
sns.scatterplot(x='POS', y='QUAL', data=gene_variants)
plt.title('Variants in Specific Gene Region')
plt.xlabel('Position')
plt.ylabel('Quality Score')
plt.show()