Integrating miRNA-Target Interactions

The miRTarBase database directly provides experimentally determined miRNA-target interactions (MTIs) that can be integrated in BEL.



In [12]:

    
import time
import sys

from bio2bel_mirtarbase import Manager
from pybel import BELGraph
from pybel.dsl import *
from pybel_tools.visualization import to_jupyter



In [13]:

    
print(sys.version)









    



3.6.3 (default, Oct  9 2017, 09:47:56) 
[GCC 4.2.1 Compatible Apple LLVM 9.0.0 (clang-900.0.37)]



In [10]:

    
time.asctime()









    Out[10]:





'Mon Nov 27 16:34:55 2017'

Like all Bio2BEL packages that handle databases, there is a class Manager available from the top-level of the package that makes very sensible defaults for where to store the data and load connection configuration for more advanced users.



In [2]:

    
manager = Manager()

A simple graph is constructed with two RNA nodes. In general, inferring the central dogma on a graph that contains many proteins could provide additonal insight in case their corresponding RNAs are not included.



In [3]:

    
graph = BELGraph()
graph.add_node_from_data(rna(namespace='HGNC', name='APP'))
graph.add_node_from_data(rna(namespace='HGNC', name='DDX21'))









    Out[3]:





('RNA', 'HGNC', 'DDX21')

The function Manager.enrich_rnas gets all RNA nodes and adds their miRNA controllers.



In [4]:

    
%%time
manager.enrich_rnas(graph)









    



CPU times: user 3.65 s, sys: 1.14 s, total: 4.79 s
Wall time: 4.81 s



In [5]:

    
to_jupyter(graph)









    Out[5]:



In [6]:

    
graph = BELGraph()
graph.add_node_from_data(mirna(namespace='MIRTARBASE', name='hsa-miR-16-5p'))
graph.add_node_from_data(mirna(namespace='MIRTARBASE', name='hsa-miR-548e-3p'))









    Out[6]:





('miRNA', 'MIRTARBASE', 'hsa-miR-548e-3p')

The function Manager.enrich_mirnas gets all miRNA nodes and adds their RNA targets. Because this sometimes results in incredibly large numbers of connections being added, future improvements will allow for selection of the least promiscuous targets and highest confidence interactions based on their experiments.



In [7]:

    
%%time
manager.enrich_mirnas(graph)









    



CPU times: user 49.1 s, sys: 16.2 s, total: 1min 5s
Wall time: 1min 5s



In [8]:

    
to_jupyter(graph)









    Out[8]: