convert json to gff

Question

[mail]

I have annotated a genome using maker and the json output can be used with jbrowse. All data are available on my local PC. So far so good, but now I would like to export the annotated regions of the json tree into gff or fasta files to use it e.g. for a local blast search (via viroblast). I now, that I can export a single file, but I've got more than 3000 annotated genes. Unfortunately, I found several tools to get json format from many other file formats but not the other way round.

And please keep in mind, that - unfortunately - I am a newby to both: genome annotation and unix based server systems ...

Is there any simple solution?

Thomas

[Alex Reynolds]

Can you post a snippet of the JSON?

Answer 1

[Felix Francis]

You could use the following json2gff python script:

http://pythonhosted.org/mgkit/scripts/json2gff.html

usage: json2gff mongodb [-h] [-v | --quiet] [--cite] [--manual] [--version] [input_file] [output_file]

[GenoMax]

This bit from the page you linked:

Annotations converted into MongoDB records with get-gff-info mongodb can be converted back into a GFF file using this command.

That makes me think this script may work for a specific application. We shall see if @mail is able to make it work.

Answer 2

[Alex Reynolds]

It depends on the structure of your JSON, but you can do this type of conversion quite easily in Python. Here's a generic, basic example:

	#!/usr/bin/env python
	import sys
	import json
	test_json_str = '''
	[
	{"name":"FOO", "chr":"chrN", "start":100, "stop":200},
	{"name":"BAR", "chr":"chrN", "start":300, "stop":400},
	{"name":"BAZ", "chr":"chrN", "start":700, "stop":750}
	]
	'''.strip()
	test_json_obj = json.loads(test_json_str)
	for gene_obj in test_json_obj:
	line = '\t'.join([gene_obj["chr"],
	str(gene_obj["start"]),
	str(gene_obj["stop"]),
	gene_obj["name"]])
	sys.stdout.write("%s\n" % line)

view raw json2tdv.py hosted with ❤ by GitHub

You just need to work with the structure of your JSON object to figure out how to loop through your array of annotation objects. You then dereference the keys in each annotation object, writing out GFF fields in an order that follows the format's specification. The example above turns the output into three-column BED, but you can simply reorder fields to get a different format.