pyani.pyani_classify module

Module providing functions to generate clusters/species hypotheses.

class pyani.pyani_classify.Cliquesinfo[source]

Bases: tuple

Summary of clique structure.

all_k_complete

Alias for field number 2

n_nodes

Alias for field number 0

n_subgraphs

Alias for field number 1

pyani.pyani_classify.all_components_k_complete(graph: networkx.classes.graph.Graph) → bool[source]

Return True if all components in passed graph are k-complete.

Parameters:graph – NetworkX Graph object
pyani.pyani_classify.analyse_cliques(graph: networkx.classes.graph.Graph) → pyani.pyani_classify.Cliquesinfo[source]

Return Cliquesinfo NamedTuple describing clique data for a graph.

Parameters:graph – NetworkX Graph object
pyani.pyani_classify.build_graph_from_results(results, label_dict: Dict[int, str], cov_min: float = 0, id_min: float = 0) → networkx.classes.graph.Graph[source]

Return undirected graph representing the passed ANIResults object.

The passed ANIResults object is converted to an undirected graph where nodes on the graph represent genomes, and edges represent pairwise comparisons having the minimum coverage and identity indicated.

Parameters:
  • results –
    • Run object from pyani_orm
  • label_dict – dictionary of genome labels for result matrices the dict is keyed by the index/column values for the results matrices
  • cov_min –
    • minimum coverage for an edge
  • id_min –
    • minimum identity for an edge
pyani.pyani_classify.k_complete_component_status(graph: networkx.classes.graph.Graph) → List[bool][source]

Return list of Booleans of whether connected components of the graph are k-complete.

Parameters:graph – NetworkX Graph object

For each component in the passed graph, a list of Booleans is calculated, representing whether each node has property P: the degree of the node is equal to the number of nodes in that component, minus 1.

The all() gives a Boolean indicating whether all nodes in that component have property P.

pyani.pyani_classify.remove_low_weight_edges(graph: networkx.classes.graph.Graph, threshold: float, attribute: str = 'identity') → Tuple[networkx.classes.graph.Graph, List[T]][source]

Return graph and edgelist where edges having weight < threshold are removed.

Parameters:
  • graph – NetworkX Graph
  • threshold – float, minimum edge weight
  • attribute – String, attribute to use as weight