#!/usr/bin/env python
# coding: utf-8

# In[1]:


get_ipython().run_line_magic('matplotlib', 'inline')
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt


# In[3]:


get_ipython().system('head -n 35 GSE5859_series_matrix.txt')


# In[4]:


df = pd.read_table('GSE5859_series_matrix.txt', skiprows=34, low_memory=False)
df


# In[5]:


matrix  = df.iloc[40:]
new_columns = [a for a in matrix.columns]


# In[6]:


new_columns[0] = 'Probe'


# In[7]:


matrix.columns = new_columns


# In[8]:


matrix.head()


# In[9]:


matrix.reset_index().drop('index', 1)


# In[40]:


ethnicity = pd.read_csv('ethnicity.csv')


# In[43]:


ethnicity['sample_id'] = ethnicity['filename'].apply(lambda x: x[:8])
ethnicity


# In[ ]: