Convert Gencode GFF3 to GeneSearch widget JSON

gff3_gene_to_genesearch_json.py

#!/usr/bin/env python

import sys
import json

genes = []

for line in sys.stdin:
  elems = line.rstrip().split('\t')
  chromosome = elems[0]
  start = int(elems[3])
  end = int(elems[4])
  length = end - start
  md = elems[8].split(';')
  hgnc = list(filter(lambda x: x.startswith('gene_name='), md))[0].split('=')[1]
  type = list(filter(lambda x: x.startswith('gene_type='), md))[0].split('=')[1]
  posneg = elems[6]
  gene = {
    'chromosome': chromosome,
    'start': start,
    'end': end,
    'length': length,
    'hgnc': hgnc,
    'type': type,
    'posneg': posneg
  }
  genes.append(gene)
  
sys.stdout.write(json.dumps(genes, separators=(',', ':')))

From text GFF3:

$ ./gff3_gene_to_genesearch_json.py < <(awk -vFS="\t" -vOFS="\t" '($3 == "gene")' gencode.gff3) > genesearch.json

Or from compressed records:

$ ./gff3_gene_to_genesearch_json.py < <(gunzip -c gencode.gff3.gz | awk -vFS="\t" -vOFS="\t" '($3 == "gene")') > genesearch.json

Or from the web:

$ ./gff3_gene_to_genesearch_json.py < <(wget -qO- https://url/path/to/gencode.gff3.gz | gunzip -c | awk -vFS="\t" -vOFS="\t" '($3 == "gene")') > genesearch.json