wderekjones / protein_binding Goto Github PK

Need to add safeguards for the case when molecular descriptors are not specified, when summary features are not specified, etc;

add support for writing an aggregate output file.

An open problem is to write a method that merges all of the data frames generated from reading each of the files into a single dataframe.

write the output data to hdf5 files, one for each molecular descriptor. This will reduce the burden of memory usage when creating the dataset.

An open problem is to add a method that writes the output to a .csv/etc; and to allow the user to specify specific features to include in the output.

An open problem is to write a method to merge entries from the files containing molecular descriptors to their respective drug-protein compound counterpart entries from the docking and mmgbsa results.

Need to write a function to load drug features. Need to update the driver function to merge these to the output dataframe.

currently read_input_files() drops rows that don't match upon both proteinName and moleculeName. Instead of dropping these, add null values when both fields don't match.

Include summary features obtained from docking process and parse protein-complex labels