I would like some help/ advice on how to parse this file for Gene ontology (.obo)
I am working to create a visualisation in D3, and need to create a "tree" file, in the JSON format -
{
"name": "flare",
"description": "flare",
"children": [
{
"name": "analytic",
"description": "analytics",
"children": [
{
"name": "cluster",
"description": "cluster",
"children": [
{"name": "Agglomer", "description": "AgglomerativeCluster", "size": 3938},
{"name": "Communit", "description": "CommunityStructure", "size": 3812},
{"name": "Hierarch", "description": "HierarchicalCluster", "size": 6714},
{"name": "MergeEdg", "description": "MergeEdge", "size": 743}
]
}, etc..
This format seems fairly easy to replicate in a dictionary in python, with 3 fields for each entry: name, description, and children[].
My probelm here is actually HOW to extract the data. The file linked above has "objects" structured as:
[Term]
id: GO:0000001
name: mitochondrion inheritance
namespace: biological_process
def: "The distribution of mitochondria, including the mitochondrial genome, into daughter cells after mitosis or meiosis, mediated by interactions between mitochondria and the cytoskeleton." [GOC:mcc, PMID:10873824, PMID:11389764]
synonym: "mitochondrial inheritance" EXACT []
is_a: GO:0048308 ! organelle inheritance
is_a: GO:0048311 ! mitochondrion distribution
Where I will need the id, is_a and name fields. I have tried using python to parse this, but I cant seem to find a way to locate each object.
Any ideas?
Here's a fairly simple way to parse the objects in your '.obo' file. It saves the object data into a dict with the id as the key and the name and is_a data saved in a list. Then it pretty-prints it using the standard json module's .dumps function.
For testing purposes, I used a truncated version of the file in your link that only includes up to id: GO:0000006.
This code ignores any objects that contain the is_obsolete field. It also removes the description info from the is_a fields; I figured you probably wanted that, but it's easy enough to disable that functionality.
#!/usr/bin/env python
''' Parse object data from a .obo file
From http://stackoverflow.com/q/32989776/4014959
Written by PM 2Ring 2015.10.07
'''
from __future__ import print_function, division
import json
from collections import defaultdict
fname = "go-basic.obo"
term_head = "[Term]"
#Keep the desired object data here
all_objects = {}
def add_object(d):
#print(json.dumps(d, indent = 4) + '\n')
#Ignore obsolete objects
if "is_obsolete" in d:
return
#Gather desired data into a single list,
# and store it in the main all_objects dict
key = d["id"][0]
is_a = d["is_a"]
#Remove the next line if you want to keep the is_a description info
is_a = [s.partition(' ! ')[0] for s in is_a]
all_objects[key] = d["name"] + is_a
#A temporary dict to hold object data
current = defaultdict(list)
with open(fname) as f:
#Skip header data
for line in f:
if line.rstrip() == term_head:
break
for line in f:
line = line.rstrip()
if not line:
#ignore blank lines
continue
if line == term_head:
#end of term
add_object(current)
current = defaultdict(list)
else:
#accumulate object data into current
key, _, val = line.partition(": ")
current[key].append(val)
if current:
add_object(current)
print("\nall_objects =")
print(json.dumps(all_objects, indent = 4, sort_keys=True))
output
all_objects =
{
"GO:0000001": [
"mitochondrion inheritance",
"GO:0048308",
"GO:0048311"
],
"GO:0000002": [
"mitochondrial genome maintenance",
"GO:0007005"
],
"GO:0000003": [
"reproduction",
"GO:0008150"
],
"GO:0000006": [
"high-affinity zinc uptake transmembrane transporter activity",
"GO:0005385"
]
}