add gimaps scripts

gimaps/data.py

@@ -0,0 +1,69 @@
+import pandas as pd
+
+
+def read_table_file(filepath):
+    if '.xlsx' in filepath:
+        df = pd.read_excel(filepath, index_col=0)
+    if '.txt' in filepath:
+        df = pd.read_csv(filepath, sep='\t', index_col=0)
+        df.columns = df.index
+    return df
+
+
+def get_matrix(df):
+    out = df.set_index(['target1', 'target2'])['value'].unstack()
+    out.columns.name = None
+    out.index.name = None
+
+    return out
+
+
+def loadGImapTableHorlbeck_et_al(datasets_path):
+    """Read GI map data.py from Horlbeck et al, _Cell_ (2018) 'Mapping the Genetic Landscape of Human Cells'
+    https://doi.org/10.1016/j.cell.2018.06.010
+    Table S5
+    """
+    filepath = f'{datasets_path}/Horlbeck_et_al_Cell_2018/mmc5.xlsx'
+
+    tabs = pd.ExcelFile(filepath).sheet_names
+
+    print("Load Table S5 from Horlbeck et al, _Cell_ (2018)")
+    print("https://doi.org/10.1016/j.cell.2018.06.010\n")
+
+    print("Sheet Names:\n")
+    for i, t in enumerate(tabs): print(f'{i + 1}. {t}')
+
+    sheet_name = 'gene GI scores and correlations'
+    print(f'\nload "{sheet_name}" ...')
+
+    data = pd.read_excel(filepath, index_col=[0, 1], header=[0, 1, 2], sheet_name=sheet_name)
+
+    return data
+
+
+def annGImapTable(df):
+    """annotate targets in GImap table
+    col 1: target 1
+    col 2: target 2
+    col 3: GI score
+    """
+    df.insert(0, 'targetType', '')
+
+    df.loc[(df.target1.eq('non-targeting') & df.target2.eq('non-targeting')), df.columns.isin(
+        ['targetType'])] = 'ctrl-ctrl'
+    df.loc[(df.target1.eq('non-targeting') & ~df.target2.eq('non-targeting')), df.columns.isin(
+        ['targetType'])] = 'ctrl-geneB'
+    df.loc[(~df.target1.eq('non-targeting') & df.target2.eq('non-targeting')), df.columns.isin(
+        ['targetType'])] = 'geneA-ctrl'
+    df.loc[
+        (~df.target1.eq('non-targeting') & ~df.target2.eq('non-targeting')) &
+        (df.target1 == df.target2),
+        df.columns.isin(['targetType'])
+    ] = 'gene-gene'
+    df.loc[
+        (~df.target1.eq('non-targeting') & ~df.target2.eq('non-targeting')) &
+        (df.target1 != df.target2),
+        df.columns.isin(['targetType'])
+    ] = 'geneA-geneB'
+
+    return df