Mercurial > repos > in_silico > cravat_vcf_convert
changeset 24:e46db72f5b74 draft
Uploaded
| author | in_silico |
|---|---|
| date | Wed, 18 Jul 2018 10:25:27 -0400 |
| parents | f8703ba30180 |
| children | f447e525d3b2 |
| files | cravat_annotate/cravat_annotate.py cravat_annotate/cravat_annotate.xml cravat_convert/__pycache__/base_converter.cpython-36.pyc cravat_convert/__pycache__/vcf_converter.cpython-36.pyc cravat_convert/base_converter.py cravat_convert/cravat_convert.py cravat_convert/cravat_convert.xml cravat_convert/vcf_converter.py |
| diffstat | 8 files changed, 365 insertions(+), 153 deletions(-) [+] |
line wrap: on
line diff
--- a/cravat_annotate/cravat_annotate.py Wed Jul 18 10:25:17 2018 -0400 +++ /dev/null Thu Jan 01 00:00:00 1970 +0000 @@ -1,128 +0,0 @@ -import requests -import json -import sys -import re -import argparse -from __future__ import print_function - - -# The endpoint that CravatQuerys are submitted to -endpoint = 'http://www.cravat.us/CRAVAT/rest/service/query' - - -# newline and delimiter values used in the output file -delimiter = "\t" -newline = "\n" - - -# Defualt indices for intepretting a cravat file's row of data in to a CravatQuery -cr_mapping = { - 'chromosome': 1, - 'position': 2, - 'strand': 3, - 'reference': 4, - 'alternate': 5 -} - - -# The neccessary attributes neeeded to submit a query. -query_keys = [ - 'chromosome', 'position', 'strand', 'reference', 'alternate' -] - - -# Expected response keys from server. Ordered in list so that galaxy output has uniform column ordering run-to-run. -# If cravat server returns additional keys, they are appended to and included in output. -ordered_keys = [ - "Chromosome", "Position", "Strand", "Reference base(s)", "Alternate base(s)", - "HUGO symbol", "S.O. transcript", "Sequence ontology protein change", "Sequence ontology", - "S.O. all transcripts", "gnomAD AF", "gnomAD AF (African)", "gnomAD AF (Amrican)", - "gnomAD AF (Ashkenazi Jewish)", "gnomAD AF (East Asian)", "gnomAD AF (Finnish)", - "gnomAD AF (Non-Finnish European)", "gnomAD AF (Other)", "gnomAD AF (South Asian)", - "1000 Genomes AF", "ESP6500 AF (average)", "ESP6500 AF (European American)", - "ESP6500 AF (African American)", "COSMIC transcript", "COSMIC protein change", - "COSMIC variant count [exact nucleotide change]", "cosmic_site_nt", "CGL driver class", - "TARGET", "dbSNP", "cgc_role", "cgc_inheritance", "cgc_tumor_type_somatic", - "cgc_tumor_type_germline", "ClinVar", "ClinVar disease identifier", "ClinVar XRef", - "GWAS Phenotype (GRASP)", "GWAS PMID (GRASP)", "Protein 3D variant" -] - - -def get_args(): - parser = argparse.ArgumentParser() - parser.add_argument('--input', - '-i', - required = True, - help='Input path to a cravat file for querying',) - parser.add_argument('--output', - '-o', - default = None, - help = 'Output path to write results from query') - return parser.parse_args() - - -def format_chromosome(chrom): - """ : Ensure chromosome entry is propely formatted for use as querying attribute. """ - if chrom[0:3] == 'chr': - return chrom - return 'chr' + str(chrom) - - -def get_query_string(row): - """ : From a row dict, return a query string for the Cravat server. - : The row dict is cravat headeres associated to their values of that row. - """ - return '_'.join([ row['chromosome'], row['position'], row['strand'], row['reference'], row['alternate'] ]) - - -def query(in_path, out_path): - """ : From a Cravat the file at in_path, query each line on the Cravat server. - : Write the response values to file at out_path. - """ - with open(in_path, 'r') as in_file, \ - open(out_path, 'w') as out_file: - for line in in_file: - try: - line = line.strip().split('\t') - # row is dict of cravat col headers assioted values in this line - row = { header: line[index] for header, index in cr_mapping.items() } - row['chromosome'] = format_chromosome(row['chromosome']) - query_string = get_query_string(row) - call = requests.get(endpoint, params={ 'mutation': query_string }) - if call.status_code != 200 or call.text == "": - raise requests.RequestException('Bad server response for query="{}". Respone code: "{}", Response Text: "{}"' - .format(query_string, call.status_code, call.text)) - json_response = json.loads(call.text) - # See if server returned additional json key-vals not expected in ordered_keys - for key in json_response: - if key not in ordered_keys: - ordered_keys.append(key) - # Write key in order of ordered_keys to standardize order of output columns - wrote = False - for key in ordered_keys: - if key in json_response: - val = json_response[key] - else: - val = None - # Standardize format for numeric values - try: - val = float(val) - val = format(val, ".4f") - except: - pass - if wrote: - out_file.write(delimiter) - out_file.write(str(val)) - wrote = True - out_file.write(newline) - except Exception as e: - print(e, file=sys.stderr) - continue - - -if __name__ == "__main__": - cli_args = get_args() - if cli_args.output == None: - base, _ = os.path.split(cli_args.input) - cli_args.output = os.path.join(base, "cravat_converted.txt") - query(cli_args.input, cli_args.output)
--- a/cravat_annotate/cravat_annotate.xml Wed Jul 18 10:25:17 2018 -0400 +++ /dev/null Thu Jan 01 00:00:00 1970 +0000 @@ -1,25 +0,0 @@ -<tool id="cravat_query" name="CRAVAT Query" version="1.0.0"> - <description>Queries CRAVAT for cancer annotation</description> - <command interpreter="python">cravat_annotate.py -i $input -o $output</command> - - <inputs> - <param format="tabular" name="input" type="data" label="Source file"/> - </inputs> - - <outputs> - <data format="tabular" name="output" /> - </outputs> - - <tests> - <test> - <param name="input" value="input_call.txt"/> - <output name="output" file="Galaxy23-[CRAVAT_Query_on_data_22].tabular"/> - </test> - </tests> - - <help> - This tool queries CRAVAT for cancer annotation. - </help> - -</tool> -
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/cravat_convert/base_converter.py Wed Jul 18 10:25:27 2018 -0400 @@ -0,0 +1,22 @@ +class BaseConverter(object): + def __init__(self): + self.format_name = None + def check_format(self,*args,**kwargs): + err_msg = 'Converter for %s format has no method check_format' %\ + self.format_name + raise NotImplementedError(err_msg) + def setup(self,*args,**kwargs): + err_msg = 'Converter for %s format has no method setup' %\ + self.format_name + raise NotImplementedError(err_msg) + def convert_line(self,*args,**kwargs): + err_msg = 'Converter for %s format has no method convert_line' %\ + self.format_name + raise NotImplementedError(err_msg) + + +class BadFormatError(Exception): + def __init__(self, message, errors=None): + super(BadFormatError, self).__init__(message) + # Support for custom error codes, if added later + self.errors = errors \ No newline at end of file
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/cravat_convert/cravat_convert.py Wed Jul 18 10:25:27 2018 -0400 @@ -0,0 +1,80 @@ +import os +import argparse +from vcf_converter import CravatConverter +from __future__ import print_function + +def get_vcf_mapping(): + """ : VCF Headers mapped to their index position in a row of VCF values. + : These are only the mandatory columns, per the VCF spec. + """ + return { + 'CHROM': 0, + 'POS': 1, + 'ID': 2, + 'REF': 3, + 'ALT': 4, + 'QUAL': 5, + 'FILTER': 6, + 'INFO': 7 + } + + +def get_args(): + parser = argparse.ArgumentParser() + parser.add_argument('--input', + '-i', + required = True, + help='Input path to a VCF file for conversion',) + parser.add_argument('--output', + '-o', + default = None, + help = 'Output path to write the cravat file to') + return parser.parse_args() + + +def convert(in_path, out_path=None, cr_sep='\t', cr_newline='\n'): + """ : Convert a VCF file to a Cravat file. + : Arguments: + : in_path: <str> path to input vcf file + : out_path: <str> path to output cravat file. Will defualt to cravat_converted.txt in the input directory. + : cr_sep: <str> the value delimiter for the output cravat file. Default value of '\\t'. + : out_newline: <str> the newline delimiter in the output cravat file. Default of '\\n' + """ + if not out_path: + base, _ = os.path.split(in_path) + out_path = os.path.join(base, "cravat_converted.txt") + + with open(in_path, 'r') as in_file, \ + open(out_path, 'w') as out_file: + + # cr_count will be used to generate the 'TR' field of the cravat rows (first header) + cr_count = 0 + # VCF lines are always assumed to be '+' strand, as VCF doesn't specify that attribute + strand = '+' + # VCF converter. Adjusts position, reference, and alternate for Cravat formatting. + converter = CravatConverter() + # A dictionary of mandatory vcf headers mapped to their row indices + vcf_mapping = get_vcf_mapping() + + for line in in_file: + if line.startswith("#"): + continue + line = line.strip().split() + # row is dict of VCF headers mapped to corresponding values of this line + row = { header: line[index] for header, index in vcf_mapping.items() } + for alt in row["ALT"].split(","): + new_pos, new_ref, new_alt = converter.extract_vcf_variant(strand, row["POS"], row["REF"], alt) + new_pos, new_ref, new_alt = str(new_pos), str(new_ref), str(new_alt) + cr_line = cr_sep.join([ + 'TR' + str(cr_count), row['CHROM'], new_pos, strand, new_ref, new_alt, row['ID'] + ]) + out_file.write(cr_line + cr_newline) + cr_count += 1 + + +if __name__ == "__main__": + cli_args = get_args() + if cli_args.output == None: + base, _ = os.path.split(cli_args.input) + cli_args.output = os.path.join(base, "cravat_converted.txt") + convert(in_path = cli_args.input, out_path = cli_args.output)
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/cravat_convert/cravat_convert.xml Wed Jul 18 10:25:27 2018 -0400 @@ -0,0 +1,20 @@ +<tool id="cravat_convert" name="CRAVAT Convert" version="1.0.0"> + <description>Converts a VCF format file to a Cravat format file</description> + <command interpreter="python">cravat_convert.py -i $input -o $output</command> + + <inputs> + <param format="tabular" name="input" type="data" label="Source file"/> + </inputs> + + <outputs> + <data format="tabular" name="output" /> + </outputs> + + <!-- <tests></tests> --> + + <help> + Converts a VCF format file to a Cravat format file + </help> + +</tool> +
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/cravat_convert/vcf_converter.py Wed Jul 18 10:25:27 2018 -0400 @@ -0,0 +1,243 @@ +""" +A module originally obtained from the cravat package. Modified to use in the vcf +converter galaxy tool. + + +Register of changes made (Chris Jacoby): + 1) Changed imports as galaxy tool won't have access to complete cravat python package + 2) Defined BadFormatError in BaseConverted file, as I didn't have the BadFormatError module +""" + +from base_converter import BaseConverter, BadFormatError +import re + +class CravatConverter(BaseConverter): + + def __init__(self): + self.format_name = 'vcf' + self.samples = [] + self.var_counter = 0 + self.addl_cols = [{'name':'phred', + 'title':'Phred', + 'type':'string'}, + {'name':'filter', + 'title':'VCF filter', + 'type':'string'}, + {'name':'zygosity', + 'title':'Zygosity', + 'type':'string'}, + {'name':'alt_reads', + 'title':'Alternate reads', + 'type':'int'}, + {'name':'tot_reads', + 'title':'Total reads', + 'type':'int'}, + {'name':'af', + 'title':'Variant allele frequency', + 'type':'float'}] + + def check_format(self, f): + return f.readline().startswith('##fileformat=VCF') + + def setup(self, f): + + vcf_line_no = 0 + for line in f: + vcf_line_no += 1 + if len(line) < 6: + continue + if line[:6] == '#CHROM': + toks = re.split('\s+', line.rstrip()) + if len(toks) > 8: + self.samples = toks[9:] + break + + def convert_line(self, l): + if l.startswith('#'): return None + self.var_counter += 1 + toks = l.strip('\r\n').split('\t') + all_wdicts = [] + if len(toks) < 8: + raise BadFormatError('Wrong VCF format') + [chrom, pos, tag, ref, alts, qual, filter, info] = toks[:8] + if tag == '': + raise BadFormatError('ID column is blank') + elif tag == '.': + tag = 'VAR' + str(self.var_counter) + if chrom[:3] != 'chr': + chrom = 'chr' + chrom + alts = alts.split(',') + len_alts = len(alts) + if len(toks) == 8: + for altno in range(len_alts): + wdict = None + alt = alts[altno] + newpos, newref, newalt = self.extract_vcf_variant('+', pos, ref, alt) + wdict = {'tags':tag, + 'chrom':chrom, + 'pos':newpos, + 'ref_base':newref, + 'alt_base':newalt, + 'sample_id':'no_sample', + 'phred': qual, + 'filter': filter} + all_wdicts.append(wdict) + elif len(toks) > 8: + sample_datas = toks[9:] + genotype_fields = {} + genotype_field_no = 0 + for genotype_field in toks[8].split(':'): + genotype_fields[genotype_field] = genotype_field_no + genotype_field_no += 1 + if not ('GT' in genotype_fields): + raise BadFormatError('No GT Field') + gt_field_no = genotype_fields['GT'] + for sample_no in range(len(sample_datas)): + sample = self.samples[sample_no] + sample_data = sample_datas[sample_no].split(':') + gts = {} + for gt in sample_data[gt_field_no].replace('/', '|').split('|'): + if gt == '.': + continue + else: + gts[int(gt)] = True + for gt in sorted(gts.keys()): + wdict = None + if gt == 0: + continue + else: + alt = alts[gt - 1] + newpos, newref, newalt = self.extract_vcf_variant('+', pos, ref, alt) + zyg = self.homo_hetro(sample_data[gt_field_no]) + depth, alt_reads, af = self.extract_read_info(sample_data, gt, gts, genotype_fields) + + wdict = {'tags':tag, + 'chrom':chrom, + 'pos':newpos, + 'ref_base':newref, + 'alt_base':newalt, + 'sample_id':sample, + 'phred': qual, + 'filter': filter, + 'zygosity': zyg, + 'tot_reads': depth, + 'alt_reads': alt_reads, + 'af': af, + } + all_wdicts.append(wdict) + return all_wdicts + + #The vcf genotype string has a call for each allele separated by '\' or '/' + #If the call is the same for all allels, return 'hom' otherwise 'het' + def homo_hetro(self, gt_str): + if '.' in gt_str: + return ''; + + gts = gt_str.strip().replace('/', '|').split('|') + for gt in gts: + if gt != gts[0]: + return 'het' + return 'hom' + + #Extract read depth, allele count, and allele frequency from optional VCR information + def extract_read_info (self, sample_data, gt, gts, genotype_fields): + depth = '' + alt_reads = '' + ref_reads = '' + af = '' + + #AD contains 2 values usually ref count and alt count unless there are + #multiple alts then it will have alt 1 then alt 2. + if 'AD' in genotype_fields and genotype_fields['AD'] <= len(sample_data): + if 0 in gts.keys(): + #if part of the genotype is reference, then AD will have #ref reads, #alt reads + ref_reads = sample_data[genotype_fields['AD']].split(',')[0] + alt_reads = sample_data[genotype_fields['AD']].split(',')[1] + elif gt == max(gts.keys()): + #if geontype has multiple alt bases, then AD will have #alt1 reads, #alt2 reads + alt_reads = sample_data[genotype_fields['AD']].split(',')[1] + else: + alt_reads = sample_data[genotype_fields['AD']].split(',')[0] + + if 'DP' in genotype_fields and genotype_fields['DP'] <= len(sample_data): + depth = sample_data[genotype_fields['DP']] + elif alt_reads != '' and ref_reads != '': + #if DP is not present but we have alt and ref reads count, dp = ref+alt + depth = int(alt_reads) + int(ref_reads) + + if 'AF' in genotype_fields and genotype_fields['AF'] <= len(sample_data): + af = float(sample_data[genotype_fields['AF']] ) + elif depth != '' and alt_reads != '': + #if AF not specified, calc it from alt and ref reads + af = float(alt_reads) / float(depth) + + return depth, alt_reads, af + + def extract_vcf_variant (self, strand, pos, ref, alt): + + reflen = len(ref) + altlen = len(alt) + + # Returns without change if same single nucleotide for ref and alt. + if reflen == 1 and altlen == 1 and ref == alt: + return pos, ref, alt + + # Trimming from the start and then the end of the sequence + # where the sequences overlap with the same nucleotides + new_ref2, new_alt2, new_pos = \ + self.trimming_vcf_input(ref, alt, pos, strand) + + if new_ref2 == '': + new_ref2 = '-' + if new_alt2 == '': + new_alt2 = '-' + + return new_pos, new_ref2, new_alt2 + + # This function looks at the ref and alt sequences and removes + # where the overlapping sequences contain the same nucleotide. + # This trims from the end first but does not remove the first nucleotide + # because based on the format of VCF input the + # first nucleotide of the ref and alt sequence occur + # at the position specified. + # End removed first, not the first nucleotide + # Front removed and position changed + def trimming_vcf_input(self, ref, alt, pos, strand): + pos = int(pos) + reflen = len(ref) + altlen = len(alt) + minlen = min(reflen, altlen) + new_ref = ref + new_alt = alt + new_pos = pos + # Trims from the end. Except don't remove the first nucleotide. + # 1:6530968 CTCA -> GTCTCA becomes C -> GTC. + for nt_pos in range(0, minlen - 1): + if ref[reflen - nt_pos - 1] == alt[altlen - nt_pos - 1]: + new_ref = ref[:reflen - nt_pos - 1] + new_alt = alt[:altlen - nt_pos - 1] + else: + break + new_ref_len = len(new_ref) + new_alt_len = len(new_alt) + minlen = min(new_ref_len, new_alt_len) + new_ref2 = new_ref + new_alt2 = new_alt + # Trims from the start. 1:6530968 G -> GT becomes 1:6530969 - -> T. + for nt_pos in range(0, minlen): + if new_ref[nt_pos] == new_alt[nt_pos]: + if strand == '+': + new_pos += 1 + elif strand == '-': + new_pos -= 1 + new_ref2 = new_ref[nt_pos + 1:] + new_alt2 = new_alt[nt_pos + 1:] + else: + new_ref2 = new_ref[nt_pos:] + new_alt2 = new_alt[nt_pos:] + break + return new_ref2, new_alt2, new_pos + + +if __name__ == "__main__": + c = CravatConverter() \ No newline at end of file