graphein.grn

Config

class graphein.grn.config.GRNGraphConfig(*, kwargs: dict[str, str | int | float] | None = {}, trrust_config: TRRUSTConfig | None = TRRUSTConfig(filtering_functions=None, root_dir=None, kwargs=None), regnetwork_config: RegNetworkConfig | None = RegNetworkConfig(filtering_functions=None, root_dir=None, kwargs=None))[source]

Config object for gene regulatory network graph construction.

Parameters:

  • kwargs (Dict[str, Union[str, int, float]], optional) – Keyword args for GRN graph construction

  • trrust_config (graphein.grn.config.TRRUSTConfig, optional) – Config object specifying parameters for parsing TRRUST. Defaults to default config object.

  • regnetwork_config (graphein.grn.config.RegNetworkConfig, optional) – Config object specifying parameters for parsing RegNetwork. Defaults to default config object.

model_config = {'arbitrary_types_allowed': True}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

class graphein.grn.config.RegNetworkConfig(*, filtering_functions: list[Callable] | None = None, root_dir: Path | None = None, kwargs: dict[str, str | int | float] | None = None)[source]

Config object containing parameters for parsing gene regulatory networks from RegNetwork: http://regnetworkweb.org/.

Parameters:

filtering_functions (List[Callable], optional) – List of functions to apply to the the RegNetwork dataframe prior to graph construction. Defaults to None

model_config = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

class graphein.grn.config.TRRUSTConfig(*, filtering_functions: list[Callable] | None = None, root_dir: Path | None = None, kwargs: dict[str, str | int | float] | None = None)[source]

Config object for parsing gene regulatory networks from TRRUST: https://www.grnpedia.org/trrust/

Parameters:

  • filtering_functions (List[Callable], optional) – List of functions to apply to the the TRRUST dataframe prior to graph construction. Defaults to None

  • root_dir (pathlib.Path, optional) – Specifies location of TRRUST dataset (will download to this path if not available). Defaults to None.

  • kwargs (Dict[str, Union[str, int, float]], optional)

model_config = {}

Configuration for the model, should be a dictionary conforming to [ConfigDict][pydantic.config.ConfigDict].

Graphs

Graph construction utilities for Gene Regulatory Networks.

graphein.grn.graphs.compute_grn_graph(gene_list: list[str], edge_construction_funcs: list[Callable], graph_annotation_funcs: list[Callable] | None = None, node_annotation_funcs: list[Callable] | None = None, edge_annotation_funcs: list[Callable] | None = None, config: GRNGraphConfig | None = None) → Graph[source]

Computes a Gene Regulatory Network Graph from a list of gene IDs

Parameters:

  • gene_list (List[str]) – List of gene identifiers

  • edge_construction_funcs (List[Callable]) – List of functions to construct edges with

  • graph_annotation_funcs (List[Callable], optional) – List of functions functools annotate graph metadata, defaults to None

  • node_annotation_funcs (List[Callable], optional) – List of functions to annotate node metadata, defaults to None

  • edge_annotation_funcs (List[Callable], optional) – List of functions to annotate edge metadata, defaults to None

  • config (graphein.grn.GRNGraphConfig, optional) – Config specifying additional parameters for STRING and BIOGRID, defaults to None

Returns:

nx.Graph of PPI network

Return type:

nx.Graph

graphein.grn.graphs.parse_kwargs_from_config(config: GRNGraphConfig) → GRNGraphConfig[source]

If configs for specific dataset are provided in the Global GRNGraphConfig, we update the kwargs

Parameters:

config (graphein.grn.GRNGraphConfig) – GRN graph configuration object.

Returns:

config with updated config.kwargs

Return type:

graphein.grn.GRNGraphConfig

Edges

Functions for adding edges to Gene Regulatory Networks from various sources.

graphein.grn.edges.add_interacting_genes(G: Graph, df: DataFrame, kind: str) → Graph[source]

Generic function for adding interaction edges to GRNGraph.

Parameters:

  • G (nx.Graph) – GRNGraph to populate with edges.

  • df (pd.DataFrame) – DataFrame containing edgelist.

  • kind (str) – name of interaction type.

Returns:

Graph with edges added.

Return type:

nx.Graph

graphein.grn.edges.add_regnetwork_edges(G: Graph, regnetwork_filtering_funcs: list[Callable] | None = None) → Graph[source]

Adds edges from RegNetwork to GRNGraph

Parameters:

  • G (nx.Graph) – Graph to edges to (populated with gene_id nodes).

  • kwargs – Additional parameters to pass to RegNetwork.

Returns:

nx.Graph GRNGraph with RegNetwork regulatory interactions added as edges.

graphein.grn.edges.add_trrust_edges(G: Graph, trrust_filtering_funcs: list[Callable] | None = None) → Graph[source]

Adds edges from TRRUST to GRNGraph

Parameters:

  • G (nx.Graph) – Graph to edges to (populated with gene_id nodes)

  • trrust_filtering_funcs (List[Callable], optional) – List of functions to apply to TRRUST dataframe as pre-processing prior to graph constructions. Defaults to None.

Returns:

nx.Graph GRNGraph with TRRUST regulatory interactions added as edges

Return type:

nx.Graph

Features

Node Features

Node featurisation utilities for Gene Regulatory Networks.

graphein.grn.features.node_features.add_sequence_to_nodes(n: str, d: dict[str, Any])[source]

Maps UniProt ACC to UniProt ID. Retrieves sequence from UniProt and adds it to the node

Parameters:

  • n (str) – Graph node. Unused (retained for a consistent function signature).

  • d (Dict[str, Any]) – Graph attribute dictionary.

Database Parsers

RegNetwork

Functions for parsing GRNs from RegNetwork.

graphein.grn.parse_regnetwork.RegNetwork_df(gene_list: list[str], root_dir: Path | None = None, filtering_funcs: list[Callable] | None = None) → DataFrame[source]

Generates standardised dataframe with RegNetwork protein-protein interactions, filtered according to user’s input.

Parameters:

  • gene_list (List[str]) – List of gene identifiers.

  • root_dir (pathlib.Path, optional. Default is None.) – Path object specifying the location of the RegNetwork dataset.

  • filtering_funcs (List[Callable], optional. Default is None.) – List of (optional) functions that accept a dataframe and return a dataframe. These functions are applied to the dataframe to filter entries according to the user’s input.

Returns:

Standardised dataframe with RegNetwork interactions.

Return type:

pd.DataFrame

graphein.grn.parse_regnetwork.filter_RegNetwork(df: DataFrame, funcs: list[Callable] | None = None) → DataFrame[source]

Filters results of RegNetwork call by providing a list of user-defined functions that accept a DataFrame and return a DataFrame

Parameters:

  • df – pd.DataFrame to filter

  • funcs – List of functions that carry out DataFrame processing.

Returns:

Processed DataFrame.

Return type:

pd.DataFrame

graphein.grn.parse_regnetwork.load_RegNetwork_interactions(root_dir: Path | None = None) → DataFrame[source]

Loads RegNetwork interaction datafile. Downloads the file first if not already present.

Parameters:

root_dir (pathlib.Path, optional. Default is None.) – Path object specifying the location of the RegNetwork download.

Returns:

RegNetwork interaction data as a pd.DataFrame.

Return type:

pd.DataFrame

graphein.grn.parse_regnetwork.load_RegNetwork_regulation_types(root_dir: Path | None = None) → DataFrame[source]
Loads RegNetwork regulation types. Downloads the file first if not already

present.

Parameters:

root_dir (pathlib.Path, optional. Default is None.) – Path object specifying the location of the RegNetwork download.

Returns:

RegNetwork regulation types as a pd.DataFrame.

Return type:

pd.DataFrame

graphein.grn.parse_regnetwork.parse_RegNetwork(gene_list: list[str], root_dir: Path | None = None) → DataFrame[source]

Parser for RegNetwork interactions

Parameters:

  • gene_list (List[str]) – List of gene identifiers

  • root_dir – Path object specifying the location of the RegNetwork download.

Returns:

Pandas DataFrame with the regulatory interactions between genes in the gene list.

Return type:

pd.DataFrame

graphein.grn.parse_regnetwork.standardise_RegNetwork(df: DataFrame) → DataFrame[source]

Standardises RegNetwork dataframe, e.g. puts everything into a common format

Parameters:

df (pd.DataFrame) – Source specific Pandas dataframe

Returns:

Standardised dataframe

Return type:

pd.DataFrame

TRRUST

Utilities for parsing the TRRUST database.

graphein.grn.parse_trrust.TRRUST_df(gene_list: list[str], filtering_funcs: list[Callable] | None = None) → DataFrame[source]

Generates standardised dataframe with TRRUST protein-protein interactions, filtered according to user’s input.

Parameters:

  • gene_list (List[str]) – List of gene identifiers to restrict dataframe to.

  • filtering_funcs (List[Callable]) – Functions with which to filter the dataframe.

Returns:

Standardised dataframe with TRRUST interactions.

Return type:

pd.DataFrame

graphein.grn.parse_trrust.filter_TRRUST(df: DataFrame, funcs: list[Callable] | None) → DataFrame[source]

Filters results of TRRUST call according to user kwargs.

Parameters:

  • df (pd.DataFrame) – Source specific Pandas dataframe (TRRUST) with results of the API call

  • funcs (List[Callable]) – User functions to filter the results.

Returns:

Source specific Pandas dataframe with filtered results

Return type:

pd.DataFrame

graphein.grn.parse_trrust.load_TRRUST(root_dir: Path | None = None) → DataFrame[source]

Loads the TRRUST datafile. If file not found, it is downloaded.

Parameters:

root_dir (pathlib.Path, optional) – Root directory path to either find or download TRRUST

Returns:

TRRUST database as a dataframe

Return type:

pd.DataFrame

graphein.grn.parse_trrust.parse_TRRUST(gene_list: list[str], root_dir: Path | None = None) → DataFrame[source]

Parser for TRRUST regulatory interactions. If the TRRUST dataset is not found in the specified root_dir, it is downloaded

Parameters:

  • gene_list (List[str]) – List of gene identifiers to restrict dataframe to.

  • root_dir (pathlib.Path, optional) – Root directory path to either find or download TRRUST. Defaults to None (downloads dataset to graphein/datasets/trrust)

Returns:

Pandas DataFrame with the regulatory interactions between genes in the gene list.

Return type:

pd.DataFrame

graphein.grn.parse_trrust.standardise_TRRUST(df: DataFrame) → DataFrame[source]

Filters results of TRRUST call by providing a list of user-defined functions that accept a dataframe and return a dataframe.

Parameters:

  • df (pd.DataFrame) – pd.Dataframe to filter. Must contain columns: ["g1", "g2", "regtype"].

  • funcs (List[Callable]) – List of functions that carry out dataframe processing.

Returns:

Processed DataFrame.

Return type:

pd.DataFrame