added script to summarize and annotate main results

experiments/assessing_cluster_clonality/sandbox/annotate_summary.py

@@ -0,0 +1,80 @@
+#### The script takes the merged splitting summary table and maps the number annotations back to the original names. Furthermore, it filters out the
+#### clusters that are in fact no CTC clutsers put primary tumor samples 
+
+import pandas as pd
+from pathlib import Path
+
+
+
+
+
+
+def load_node_description(sample_name, color_annotation):
+    file_path = Path('/home/jovyan/work/ctc-data/WES_experiment/splitting_summaries/') / Path(sample_name) / f"{sample_name}_samples_nodeDescription.tsv"
+    df = pd.read_csv(file_path, sep='\t', header = None)
+    return df
+
+def search_rows_with_color_annotation(df, color_annotation):
+    color_annotation = f"={color_annotation},"
+    return ",".join(df[df.iloc[:, -1].str.contains(color_annotation, na=False)].iloc[:, 0].values.tolist())
+
+def remove_primary_tumor_samples(line, master_table):
+    samples = line.split(',')
+    sample_constituents = []
+    for s in samples:
+        if not 'CTC' in master_table.loc[master_table['sample_id'] == s, 'sample_type'].to_string():
+            continue
+        else:
+            sample_constituents.append(s)
+    remaining_samples = ','.join(sample_constituents)
+    return remaining_samples
+
+
+def filter_out_non_CTC_clusters(splitting_summary, master_table):
+    splitting_summary_filtered = splitting_summary.copy(deep=True) 
+    splitting_summary_filtered['sample_names'] = splitting_summary_filtered['sample_names'].map(lambda row: remove_primary_tumor_samples(row,master_table))
+    rows_to_keep = [idx for idx, row in splitting_summary_filtered.iterrows() if len(row['sample_names'])>0]
+    splitting_summary_filtered = splitting_summary_filtered.loc[rows_to_keep,]
+
+    return splitting_summary_filtered
+
+def compute_number_of_cells(sample_names, master_table, WBC = False):
+    all_sample_names = sample_names.split(',')
+
+    if WBC:
+        cell_count_column = 'n_wbc_attached'
+    else:
+        cell_count_column = 'n_cells'
+    n_cells = 0
+    for sample in all_sample_names:
+        n_cells += int(master_table.loc[master_table['sample_id'] == sample, cell_count_column].iloc[0])
+    return n_cells
+
+
+def main():
+    splitting_summary = pd.read_csv('/home/jovyan/work/ctc-data/WES_experiment/splitting_summaries/splittingSummary_full.tsv', sep = '\t', header = 0)
+    splitting_summary['sample_names'] = splitting_summary.apply(
+            lambda row: search_rows_with_color_annotation(load_node_description(row['Sample Name'], row['Color']), row['Color']),
+            axis=1
+        )
+
+    master_table = pd.read_csv('/home/jovyan/work/CTC-SCITE/experiments/preprocessing/gDNA_sample_annotation_final.tsv', sep = '\t')
+
+    splitting_summary['n_cells'] = splitting_summary.apply(
+        lambda row: compute_number_of_cells(row['sample_names'],master_table),
+        axis = 1
+    )
+
+    splitting_summary['n_wbcs'] = splitting_summary.apply(
+        lambda row: compute_number_of_cells(row['sample_names'],master_table, WBC = True),
+        axis = 1
+    )
+
+
+    splitting_summary = filter_out_non_CTC_clusters(splitting_summary,master_table)
+
+
+    splitting_summary.to_csv('splittingSummary_full_with_sample_names.tsv', sep = '\t', index = False)
+
+if __name__ == '__main__':
+    main()