Script to use the WormBase REST API to retrieve phenotypes (from RNAi, mutants) for an input list of C.elegans genes

retrieve_phenotypeinfo_from_wormbase_for_genelist.py

import os
import sys
import requests # this is used to access json files

#====================================================================#

# use the wormbase REST API to retrieve the phenotypes (from mutants, RNAi) for a particular gene:

def retrieve_phenotypes_from_wormbase(gene):

    server = "http://rest.wormbase.org"
    ext = "/rest/field/gene/%s/phenotype" % gene

    r = requests.get(server+ext, headers={ "Content-Type" : "application/json", "Accept" : ""})

    if not r.ok:
        r.raise_for_status()
        sys.exit()

    decoded = r.json()
    # print(decoded)

    # based on looking at the example http://rest.wormbase.org/rest/field/gene/WBGene00000079/phenotype
    results = list() # a list of the phenotype info. results
    phenotypes = decoded["phenotype"]
    phenotypes = phenotypes["data"]
    for phenotype in phenotypes:
        phenotype = phenotype["phenotype"]
        phenotype_id = phenotype["id"]
        label = phenotype["label"] 
        result = (phenotype_id, label)
        results.append(result)

    return results

#====================================================================#

# read in the input list of C.elegans genes of interest:

def read_input_list_of_genes(input_genelist_file, output_file):

    cnt = 0 
    # open the output file:
    with open(output_file, 'w') as f:
        # read in the list of C.elegans genes:
        inputfileObj = open(input_genelist_file, "r")
        for line in inputfileObj:
            line = line.rstrip()
            temp = line.split()
            # 1   WBGene00000079   
            gene = temp[1] # e.g. WBGene00000079 
            cnt += 1
            # get the phenotypes for this gene:
            print(cnt,"Finding the phenotypes for gene",gene)
            results = retrieve_phenotypes_from_wormbase(gene)
            # write the phenotypes to the output file:
            for result in results:
                (phenotype_id, label) = result
                output_line = "%s\t%s\t%s\n" % (gene, phenotype_id, label)
                f.write(output_line)
            #xxx what about if no phenotype?
        inputfileObj.close()

#====================================================================#

def main():

    # check the command-line arguments:
    if len(sys.argv) != 3 or os.path.exists(sys.argv[1]) == False:
        print("Usage: %s input_genelist_file output_file" % sys.argv[0])
        sys.exit(1)
    input_genelist_file = sys.argv[1] # input file with a list of C.elegans genes of interest
    output_file = sys.argv[2] 

    # read in the input list of C.elegans genes of interest:
    print("Reading in gene list...")
    read_input_list_of_genes(input_genelist_file, output_file) 

    print("FINISHED\n")

#====================================================================#

if __name__=="__main__":
    main()

#====================================================================#