Merge pull request SouthernBio#16 from SouthernBio/feature/zscore

Functions for analyzing dataframes

Pipfile

@@ -8,6 +8,7 @@ numpy = "*"
 pandas = "*"
 matplotlib = "*"
 sh = "*"
+scipy = "*"
 
 [dev-packages]
 

tools/analysis.py

@@ -0,0 +1,97 @@
+import pandas as pd
+from scipy.stats import zscore
+from fnmatch import fnmatch
+
+
+class NotADataframe(Exception):
+    """Custom exception for when a function receives a file that is not a CSV dataframe"""
+    message = "Error: the function must receive a dataframe in CSV format."
+
+    def __init__(self):
+        super().__init__(self.message)
+
+
+def normalize_conservation_rate(file_path: str) -> pd.DataFrame:
+    """Normalizes conservation rates by calculating the z-score for each observed conservation rate.
+
+    Args:
+        file_path (pd.DataFrame): Pandas dataframe containing conservation rates
+
+    Returns:
+        pd.DataFrame: Pandas dataframe with normalized conservation rates
+    """
+    if not fnmatch(file_path, '*.csv'):
+        raise NotADataframe()
+
+    dataframe = pd.read_csv(file_path, delimiter=',', header=0, index_col=0)
+
+    # Calculate z-score
+    dataframe[['NormalizedConservationRate']] = zscore(dataframe[['ConservationRate']], axis=None)
+
+    # IMPORTANT: Modify this function (or create a new one)
+    # to handle conservation rate values of zero
+
+    return dataframe
+
+
+def split_codon_pairs(file_path: str) -> pd.DataFrame:
+    """Generate two columns with each constituent codon from a codon pair
+
+    Args:
+        file_path (str): File which contains relevant 
+    Returns:
+        pd.DataFrame: Pandas dataframe with two new columns
+    """
+    if not fnmatch(file_path, '*.csv'):
+        raise NotADataframe()
+
+    dataframe = pd.read_csv(file_path, delimiter=',', header=0)
+
+    # Get codon pairs
+    codon_pairs = dataframe[dataframe.columns[0]]
+
+    # Generate lists with each constituent codon
+    first_codon, second_codon = [], []
+    for pair in codon_pairs:
+        first_codon.append(pair[:3])
+        second_codon.append(pair[3:])
+
+    # Join the lists to the dataframe
+    df_codons = pd.DataFrame({
+        'FirstCodon': pd.Series(first_codon, index=None),
+        'SecondCodon': pd.Series(second_codon, index=None)
+        })
+    dataframe = dataframe.join(df_codons)
+
+    return dataframe
+
+
+def expected_codon_pair_conservation_rate(file_path: str, codons: str) -> pd.DataFrame:
+
+    codon_information = pd.read_csv(codons, delimiter=',', header=0, index_col=0)
+    dataframe = pd.read_csv(file_path, delimiter=',', header=0, index_col=0)
+
+    # Auxiliary lists to store normalized conservation values
+    first_codon_cr, second_codon_cr, expected_product = [], [], []
+
+    for codon in dataframe['FirstCodon']:
+        for i in codon_information['Codon']:
+            if codon == i:
+                first_codon_cr.append(codon_information.filter(index=i, axis=0)[['ConservationRate']]) 
+
+    for codon in dataframe['SecondCodon']:
+        for i in codon_information['Codon']:
+            if codon == i:
+                second_codon_cr.append(codon_information.filter(index=i, axis=0)[['ConservationRate']])
+
+    for index, item in enumerate(first_codon_cr):
+        expected_product.append(first_codon_cr[index] * second_codon_cr[index])
+
+    expectedCodonPairConservationRate = pd.Series(expected_product, index=None)
+    df_expected_conservation_rate = pd.DataFrame({
+        'ExpectedCodonPairConservationRate': expectedCodonPairConservationRate
+    })
+
+    dataframe = dataframe.join(df_expected_conservation_rate)
+
+    return dataframe