rnaglib.transforms.RBPTransform

class rnaglib.transforms.RBPTransform(structures_dir, distance_threshold=5.0, protein_number_annotations=False, distances=[0.5])

Adds information at the residue level about the protein content of the environment of the residue. Two types of annotations are added: a binary feature indicating whether each residue is binding a protein according to a user-defined threshold, and features indicating the number of protein atoms in radiuses of certain distances around each residue. Since it is computationally expensive, it is by default only used during the dataset creation. The protein content annotation is further used to discard RNA residues with high protein content for certain tasks.

Parameters:

• structures_dir (Union[PathLike, str]) – Path to the directory where the structures are stored (ex. as cif files)

• distance_threshold (float) – the radius (in Angstrom) of the zone considered as the environment of the residue (default 5.0)

• protein_number_annotations (bool) – whether to add annotations regarding the number of protein atoms around each residue besides the binary annotation (default False)

• distances (list[float]) – the list of the radiuses (in Angstroms) of the balls centered around the residues within which we would like to compute the number of protein atoms

Returns:

the annotated graph, actually the graph is mutated in place

__init__(structures_dir, distance_threshold=5.0, protein_number_annotations=False, distances=[0.5])

Methods

__init__(structures_dir[, ...])
forward(rna_dict)	Application of the transform to an RNA dictionary object