network analysis of correlated mutations in proteins

in active development

performs network analysis (based on "maximal information coefficient") of correlated mutations in proteins.

• download strains from fludb.org
• run split_fasta.py to split the strain file to individual (protein) fasta files
• align using MUSCLE/MAFFT
• run deduplication script - duplicate_remover.py
• run graph_analysis_2p.py or graph_analysis_1p.py for computing MICs and creating CSV files with node and edge data for a given dataset
• run load_graph_db_from_csv.py to load neo4j from CSV files
• run create_networkx_graph_from_csv.py to create a graphml from CSV files
• import graphml into Gephi to create visualizations and perform analysis
• run create_protein_graph.py to create macro/protein level graphs and visualizations from graphml
• run create_degree_clustering_plots.py to create degree/clustering plots from graphml
• run create_plots_from_neo4j.py to create node and edge count plots from data loaded in neo4j
• run cooccurence_counts.py to perform residue coocurrence analysis
• run create_mic_histogram.py to create a histogram of MICs
• run in_out_comparision.py to perform entropy and solvent accessibility comparision of in-network and out-of-network residues