PANTHER (http://pantherdb.org/) interface

panther.py

# LICENSE: Public Domain

import requests


def panther(species, genelist, ontology='biological_process',
        correction='fdr', format='html', resource='PANTHER', timeout=15):
    """Post a over/under-representation test request to PANTHER."""

    # NOTE: The species parameter seems to be different from conventional
    # names. For example, Homo sapiens is 'HUMAN' and Yarrowia lipolytica is 'YARLI'.

    # request analysis
    url = 'http://pantherdb.org/webservices/go/overrep.jsp'
    data = {'species': species, 'input': "\n".join(genelist),
            'correction': correction, 'format': format, 'resource': resource}
    r = requests.post(url, data=data, timeout=timeout)
    r.raise_for_status()

    # request xml data containing the result
    url = 'http://pantherdb.org/tools/exportCompareToRefList.jsp'
    params = {'sortOrder': -1, 'sortList': 'upload_1', 'sortField': 'hierarchy', 'format': 'xml'}
    r = requests.get(url, params=params, cookies=r.cookies, timeout=timeout)
    r.raise_for_status()

    # decode the content and fix the encoding signature
    xml = r.content.decode('UTF-8')
    xml = xml.replace('encoding="UTF-16"', 'encoding="UTF-8"', 1)
    
    return xml


if __name__ == '__main__':
    import sys
    if len(sys.argv) != 3:
        print("Usage: panther.py <species> <input file>")
        sys.exit(1)
    species, inputfile = sys.argv[1:]
    with open(inputfile) as f:
        genelist = f.readlines()
        print(panther(species, genelist))