graphein.ppi

Config

class graphein.ppi.config.BioGridConfig(*, searchNames: bool = True, max: int = 10000, interSpeciesExcluded: bool = True, selfInteractionsExcluded: bool = False, evidenceList: str = '', includeEvidence: bool = False, searchIds: bool = True, searchSynonyms: bool = True, searchBiogridIds: bool = True, additionalIdentifierTypes: str = '', excludeGenes: bool = False, includeInteractors: bool = True, includeInteractorInteractions: bool = False, pubmedList: str = '', excludePubmeds: bool = False, htpThreshold: int = 20, throughputTag: str = 'any')

Config for specifying parameters for API calls to BIOGRID. A full description of the parameters can be found at : https://wiki.thebiogrid.org/doku.php/biogridrest

Parameters

• searchNames (bool, optional) – If ‘true’, the interactor OFFICIAL_SYMBOL will be examined for a match with the geneList.

• max (int, optional) – Number of results to fetch, defaults to 10,000

• interSpeciesExcluded (bool, optional) – If ‘true’, interactions with interactors from different species will be excluded, defaults to True

• selfInteractionsExcluded (bool, optional.) – If ‘true’, interactions with one interactor will be excluded, defaults to False

• evidenceList (str, optional) – Any interaction evidence with its Experimental System in the list will be excluded from the results unless includeEvidence is set to true., defaults to “” (empty string)

• includeEvidence (bool, optional) – If set to true, any interaction evidence with its Experimental System in the evidenceList will be included in the result, defaults to False

• searchIDs (bool, optional) – If ‘true’, the interactor ENTREZ_GENE, ORDERED LOCUS and SYSTEMATIC_NAME (orf) will be examined for a match with the geneList. Defaults to True

• searchNames – # If ‘true’, the interactor OFFICIAL_SYMBOL will be examined for a match with the geneList. Defaults to True.

• searchSynonyms (bool, optional) – If ‘true’, the interactor SYNONYMS will be examined for a match with the geneList. Defaults to True.

• searchBiogridIds – If ‘true’, the entries in ‘GENELIST’ will be compared to BIOGRID internal IDS which are provided in all Tab2 formatted files. Defaults to True

• additionalIdentifierTypes (str, optional) – Identifier types on this list are examined for a match with the geneList. Defaults to “”

• excludeGenes (bool, optional) – If ‘true’, interactions containing genes in the geneList will be excluded from the results. Defaults to False

• includeInteractors (bool, optional) – If ‘true’, in addition to interactions between genes on the geneList, interactions will also be fetched which have only one interactor on the geneList. Defaults to True

• includeInteractorInteractions (bool, optional) – # If ‘true’ interactions between the geneList’s first order interactors will be included. Defaults to False

• pubmedList (str, optional) – Interactions will be fetched whose Pubmed Id is/ is not in this list, depending on the value of excludePubmeds. Defaults to “”

• excludePubmeds (bool, optional) – If ‘false’, interactions with Pubmed ID in pubmedList will be included in the results; if ‘true’ they will be excluded. Defaults to False

• htpThreshold (int, optional) – Interactions whose Pubmed ID has more than this number of interactions will be excluded from the results. Ignored if excludePubmeds is ‘false’. Defaults to 20.

• throughputTag (str, optional) – If set to ‘low or ‘high’, only interactions with ‘Low throughput’ or ‘High throughput’ in the ‘throughput’ field will be returned. Defaults to “any”

class graphein.ppi.config.PPIGraphConfig(*, paginate: bool = True, ncbi_taxon_id: int = 9606, kwargs: Dict[str, Union[str, int, float]] = {'BIOGRID_throughputTag': 'high', 'STRING_escore': 0.2}, string_config: graphein.ppi.config.STRINGConfig = None, biogrid_config: graphein.ppi.config.BioGridConfig = None)

Config for specifying parameters for PPI Graph Construction

Parameters

• paginate (bool) – Controls whether or not to paginate API calls. Useful for large queries. Defaults to True

• ncbi_taxon_id (int) – Defaults to 9606 (human)

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

• string_config (graphein.ppi.config.STRINGConfig) – Config Object holding parameters for STRINGdb API calls. Defaults to None

• biogrid_config (graphein.ppi.config.BioGridConfig, optional) – Config Object holding parameters for BioGrid API calls. Defaults to None

class graphein.ppi.config.STRINGConfig(*, species: int = 9606, required_score: int = 50, network_type: str = 'functional', add_nodes: int = 0, show_query_node_labels: bool = 0)

Config for specifying parameters for API calls to STRINGdb. Full documentation can be found: https://string-db.org/help/api/

Parameters

• species (int, optional) – NCBI taxon identifiers, defaults to 9606 (human)

• required_score (int, optional) – Threshold of significance to include a interaction, a number between 0 and 1000 (default depends on the network)

• network_type (str, optional) – Network type: “functional” (default), “physical”

• add_nodes (int, optional) – Adds a number of proteins to the network based on their confidence score, e.g., extends the interaction neighborhood of selected proteins to desired value, defaults to 50

• show_query_node_labels (bool, optional) – When available use submitted names in the preferredName column when (0 or 1) (default:0)

Graph Construction

Graphs

Functions for constructing a PPI PPIGraphConfig from STRINGdb and BIOGRID.

graphein.ppi.graphs.compute_ppi_graph(protein_list: List[str], edge_construction_funcs: List[Callable], graph_annotation_funcs: Optional[List[Callable]] = None, node_annotation_funcs: Optional[List[Callable]] = None, edge_annotation_funcs: Optional[List[Callable]] = None, config: Optional[graphein.ppi.config.PPIGraphConfig] = None) → networkx.classes.graph.Graph

Computes a PPI Graph from a list of protein IDs. This is the core function for PPI graph construction.

Parameters

• protein_list (List[str]) – List of protein identifiers

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

• graph_annotation_funcs (List[Callable], optional) – List of functions to annotate graph metadata

• node_annotation_funcs (List[Callable], optional) – List of functions to annotate node metadata

• edge_annotation_funcs (List[Callable], optional) – List of function to annotate edge metadata

• config (PPIGraphConfig, optional) – Config object specifying additional parameters for STRING and BIOGRID API calls

Returns

nx.Graph of PPI network

Return type

nx.Graph

graphein.ppi.graphs.parse_kwargs_from_config(config: graphein.ppi.config.PPIGraphConfig) → graphein.ppi.config.PPIGraphConfig

If configs for STRING and BIOGRID are provided in the Global ~graphein.ppi.config.PPIGraphConfig, we update the kwargs

Parameters

config (PPIGraphConfig) – PPI graph configuration object.

Returns

config with updated config.kwargs

Return type

PPIGraphConfig

Edges

Functions for adding edges to a PPI Graph from parsed STRING & BIOGRID API call outputs.

graphein.ppi.edges.add_biogrid_edges(G: networkx.classes.graph.Graph, **kwargs) → networkx.classes.graph.Graph

Adds edges from the BIOGRID database (https://thebiogrid.org/) to PPI Graph.

Parameters

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

• kwargs – Additional parameters to pass to BIOGRID API calls.

Returns

nx.Graph PPIGraph with BIOGRID interactions added as edges.

Return type

nx.Graph

graphein.ppi.edges.add_interacting_proteins(G: networkx.classes.graph.Graph, df: pandas.core.frame.DataFrame, kind: str) → networkx.classes.graph.Graph

Generic function for adding interaction edges to PPI Graph. You can use this function to additional interactions using a dataframe with columns "p1" and "p2".

Parameters

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

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

• kind (str) – name of interaction type.

Returns

PPI Graph with pre-computed edges added.

Return type

nx.Graph

graphein.ppi.edges.add_string_edges(G: networkx.classes.graph.Graph, **kwargs) → networkx.classes.graph.Graph

Adds edges from STRING PPI database (https://string-db.org/) to a PPI Graph.

Parameters

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

• kwargs – Additional parameters to pass to STRING API calls.

Returns

PPI Graph with STRING interactions added as edges.

Return type

nx.Graph

Graph Features

Functions for adding metadata to PPI Graphs from STRING and BIOGRID.

graphein.ppi.graph_metadata.add_biogrid_metadata(G: networkx.classes.graph.Graph, kwargs: Dict[str, Union[str, int]]) → networkx.classes.graph.Graph

Adds interaction dataframe from BIOGRID to graph.

Parameters

• G (nx.Graph) – PPI Graph to add metadata to

• kwargs (Dict[str, Union[str, int]]) – Additional parameters for BIOGRID API call

Returns

PPIGraph with added BIOGRID interaction_df as metadata

Return type

nx.Graph

graphein.ppi.graph_metadata.add_string_biogrid_metadata(G: networkx.classes.graph.Graph, kwargs: Dict[str, Union[str, int]]) → networkx.classes.graph.Graph

Adds interaction dataframe from STRING and BIOGRID to graph.

Parameters

• G (nx.Graph) – PPIGraph to add metadata to

• kwargs (Dict[str, Union[str, int]]) – Additional parameters for STRING and BIOGRID API calls

Returns

PPIGraph with added STRING and BIOGRID interaction_df as metadata

Return type

nx.Graph

graphein.ppi.graph_metadata.add_string_metadata(G: networkx.classes.graph.Graph, kwargs: Dict[str, Union[str, int]]) → networkx.classes.graph.Graph

Adds interaction dataframe from STRING to graph.

Parameters

• G (nx.Graph) – PPI Graph to add metadata to

• kwargs (Dict[str, Union[str, int]]) – Additional parameters for STRING API call

Returns

PPIGraph with added STRING interaction_df as metadata

Return type

nx.Graph

Node Features

Functions for adding nodes features to a PPI Graph

graphein.ppi.features.node_features.add_sequence_to_nodes(n: str, d: Dict[str, Any])

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

Parameters

• n (str) – Graph node.

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

Visualisation

Contains utilities for plotting PPI NetworkX graphs.

graphein.ppi.visualisation.get_edge_trace(g: networkx.classes.graph.Graph, edge_colours: Optional[List[str]] = None) → List[plotly.graph_objs._scatter.Scatter]

Gets edge traces from PPI graph. Returns a list of traces enabling edge colours to be set individually.

Parameters

g (nx.Graph) – _description_

Returns

_description_

Return type

List[go.Scatter]

graphein.ppi.visualisation.get_node_trace(g: networkx.classes.graph.Graph, node_size_multiplier: float, node_colourscale: str = 'Viridis') → plotly.graph_objs._scatter.Scatter

Produces the node trace for the plotly plot.

Parameters

• g (nx.Graph) – PPI graph with [‘pos’] added to the nodes (eg via nx.layout function)

• node_size_multiplier (float) – Multiplier for node size. Default is 5.0.

• node_colourscale (str, optional) – Colourscale to use for the nodes, defaults to “Viridis”

Returns

Node trace for plotly plot

Return type

go.Scatter

graphein.ppi.visualisation.plot_ppi_graph(g: networkx.classes.graph.Graph, colour_edges_by: str = 'kind', with_labels: bool = True, **kwargs)

Plots a Protein-Protein Interaction Graph. Colours edges by kind.

Parameters

• g (nx.Graph) – NetworkX graph of PPI network.

• colour_edges_by – Colour edges by this attribute. Currently, only supports ‘kind’, which colours edges by the source database, by default “kind”

• with_labels (bool, optional) – Whether to show labels on nodes. Defaults to True.

graphein.ppi.visualisation.plotly_ppi_graph(g: networkx.classes.graph.Graph, layout: <module 'networkx.drawing.layout' from '/Users/arianjamasb/opt/anaconda3/envs/graphein-wip/lib/python3.8/site-packages/networkx/drawing/layout.py'> = <function circular_layout>, title: typing.Optional[str] = None, show_labels: bool = False, node_size_multiplier: float = 5.0, node_colourscale: str = 'Viridis', edge_colours: typing.Optional[typing.List[str]] = None, edge_opacity: float = 0.5, height: int = 500, width: int = 500)

Plots a PPI graph.

Parameters

• g (nx.Graph) – PPI graph

• layout (nx.layout) – Layout algorithm to use. Default is circular_layout.

• title (str, optional) – Title of the graph. Default is None.

• show_labels (bool) – If True, shows labels on nodes. Default is False.

• node_size_multiplier (float) – Multiplier for node size. Default is 5.0.

• node_colourscale (str) – Colour scale to use for node colours. Default is “Viridis”. Options: ‘Greys’ | ‘YlGnBu’ | ‘Greens’ | ‘YlOrRd’ | ‘Bluered’ | ‘RdBu’ | ‘Reds’ | ‘Blues’ | ‘Picnic’ | ‘Rainbow’ | ‘Portland’ | ‘Jet’ | ‘Hot’ | ‘Blackbody’ | ‘Earth’ | ‘Electric’ | ‘Viridis’ |

• edge_colours (List[str], optional) – List of colours (hexcode) to use for edges. Default is None (px.colours.qualitative.T10).

• edge_opacity (float) – Opacity of edges. Default is 0.5.

• height (int) – Height of the plot. Default is 500.

• width (int) – Width of the plot. Default is 500.

Returns

Plotly figure of PPI Network

Return type

go.Figure

Database Parsers

BioGrid

Functions for making and parsing API calls to BIOGRID.

graphein.ppi.parse_biogrid.BIOGRID_df(protein_list: List[str], ncbi_taxon_id: Union[int, str, List[int], List[str]], **kwargs) → pandas.core.frame.DataFrame

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

Protein_list

List of proteins (official symbol) that will be included in the PPI graph

Ncbi_taxon_id

NCBI taxonomy identifiers for the organism. 9606 corresponds to Homo Sapiens

Parameters

kwargs (Union[int, str, List[int], List[str]]) – Additional parameters to pass to BIOGRID API calls

Returns

Standardised dataframe with BIOGRID interactions

Return type

pd.DataFrame

graphein.ppi.parse_biogrid.filter_BIOGRID(df: pandas.core.frame.DataFrame, **kwargs) → pandas.core.frame.DataFrame

Filters results of the BIOGRID API call according to user kwargs.

Parameters

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

• kwargs (Dict[str, Union[str, int, List[str], List[int]]]) – User thresholds used to filter the results. The parameter names are of the form BIOGRID_<param>, where <param> is the name of the parameter. All the parameters are numerical values.

Returns

Source specific Pandas dataframe with filtered results

Return type

pd.DataFrame

graphein.ppi.parse_biogrid.params_BIOGRID(params: Dict[str, Union[str, int, List[str], List[int]]], **kwargs) → Dict[str, Union[str, int]]

Updates default parameters with user parameters for the method “interactions” of the BIOGRID API REST.

See also https://wiki.thebiogrid.org/doku.php/biogridrest :param params: Dictionary of default parameters :type params: Dict[str, Union[str, int, List[str], List[int]]] :param kwargs: User parameters for the method “network” of the BIOGRID API REST. The key must start with “BIOGRID” :type kwargs: Dict[str, Union[str, int, List[str], List[int]]] :return: Dictionary of parameters :rtype: Dict[str, Union[str, int]]

graphein.ppi.parse_biogrid.parse_BIOGRID(protein_list: List[str], ncbi_taxon_id: Union[int, str, List[int], List[str]], paginate: bool = True, **kwargs) → pandas.core.frame.DataFrame

Makes BIOGRID API call and returns a source specific Pandas dataframe.

See also [1] BIOGRID: https://wiki.thebiogrid.org/doku.php/biogridrest :param protein_list: Proteins to include in the graph :type protein_list: List[str] :param ncbi_taxon_id: NCBI taxonomy identifiers for the organism. Default is 9606 (Homo Sapiens) :type ncbi_taxon_id: Union[int, str, List[int], List[str]] :param paginate: boolean indicating whether to paginate the calls (for BIOGRID, the maximum number of rows per

call is 10000). Defaults to True

Parameters

kwargs (Dict[str, Union[str, int, List[str], List[int]]]) – Parameters of the “interactions” method of the BIOGRID API REST, used to select the results. The parameter names are of the form BIOGRID_<param>, where <param> is the name of the parameter. Information about these parameters can be found at [1].

Returns

Source specific Pandas dataframe.

Return type

pd.DataFrame

graphein.ppi.parse_biogrid.standardise_BIOGRID(df: pandas.core.frame.DataFrame) → pandas.core.frame.DataFrame

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

Parameters

df (pd.DataFrame) – Source specific Pandas dataframe

Returns

Standardised dataframe

Rtpe

pd.DataFrame

STRINGDB

Functions for making and parsing API calls to STRINGdb.

graphein.ppi.parse_stringdb.STRING_df(protein_list: List[str], ncbi_taxon_id: Union[int, str, List[int], List[str]], **kwargs) → pandas.core.frame.DataFrame

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

Parameters

• protein_list (List[str]) – List of proteins (official symbol) that will be included in the PPI graph

• ncbi_taxon_id (int) – NCBI taxonomy identifiers for the organism. 9606 corresponds to Homo Sapiens

• kwargs – Additional parameters to pass to STRING API calls

Returns

Standardised dataframe with STRING interactions

Return type

pd.DataFrame

graphein.ppi.parse_stringdb.filter_STRING(df: pandas.core.frame.DataFrame, **kwargs) → pandas.core.frame.DataFrame

Filters results of the STRING API call according to user kwargs, keeping rows where the input parameters are greater or equal than the input thresholds.

Parameters

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

• kwargs (Dict[str, Union[str, int, List[str], List[int]]]) – User thresholds used to filter the results. The parameter names are of the form STRING_<param>, where <param> is the name of the parameter. All the parameters are numerical values.

Returns

Source specific Pandas dataframe with filtered results

Return type

pd.DataFrame

graphein.ppi.parse_stringdb.params_STRING(params: Dict[str, Union[str, int, List[str], List[int]]], **kwargs) → Dict[str, Union[str, int]]

Updates default parameters with user parameters for the method “network” of the STRING API REST. See also https://string-db.org/help/api/

Parameters

• params (Dict[str, Union[str, int, List[str], List[int]]]) – Dictionary of default parameters

• kwargs (Dict[str, Union[str, int, List[str], List[int]]]) – User parameters for the method “network” of the STRING API REST. The key must start with “STRING”

Returns

Dictionary of parameters

Return type

Dict[str, Union[str, int]]

graphein.ppi.parse_stringdb.parse_STRING(protein_list: List[str], ncbi_taxon_id: Union[int, str, List[int], List[str]], **kwargs) → pandas.core.frame.DataFrame

Makes STRING API call and returns a source specific Pandas dataframe. See also [1] STRING: https://string-db.org/help/api/

Parameters

• protein_list (List[str]) – Proteins to include in the graph

• ncbi_taxon_id (int) – NCBI taxonomy identifiers for the organism. Default is 9606 (Homo Sapiens)

• kwargs (Dict[str, Union[str, int, List[str], List[int]]]) – Parameters of the “network” method of the STRING API REST, used to select the results. The parameter names are of the form STRING_<param>, where <param> is the name of the parameter. Information about these parameters can be found at [1].

Returns

Source specific Pandas dataframe.

Return type

pd.DataFrame

graphein.ppi.parse_stringdb.standardise_STRING(df: pandas.core.frame.DataFrame) → pandas.core.frame.DataFrame

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

Parameters

df (pd.DataFrame) – Source specific Pandas dataframe

Returns

Standardised dataframe

Return type

pd.DataFrame