Pathminer: A simple biological pathway mining algorithm
This package provides a simple Python API for biological pathway mining using the Biocyc (+MetaCyc) web services. It is built on the separate package biocyc which handles the actual connections.
This algorithm has successfully been used in the Pathomx software to identify gene and metabolite changes within pathways. Parameters to the function as follows:
data: a list of tuples in the format (biocyc_id, score); multiple datasets as a list of lists
target: what to mine - pathways, reactions or compartments (not supported yet)
algorithm: one of 'c', 'u', 'd', 'm', 't'
'Compound change scores for pathway': 'c'
'Compound up-regulation scores for pathway': 'u'
'Compound down-regulation scores for pathway': 'd'
'Number compounds with data per pathway': 'm'
'Pathway overall tendency': 't'
no_of_results: number of targets to return
shared: 'share' scores between pathways, to minimise influence of promiscuous metabolites
relative: 'scale' scores to the size of pathways, to reduce influence of pathway sizeSimple usage is as follows.
from biocyc import biocyc
biocyc.set_organism('meta')
from pathminer import mining
a = biocyc.get('FRUCTOSE-16-DIPHOSPHATE')
b = biocyc.get('GLC')
data = [(a,5),(b,10)]
mining(data)
[(glycolysis, 2.5), (gluconeogenesis, 2.5), (glycolysis, 2.5), (gluconeogenesis, 2.5), (glycogenolysis, 0.625)]
